-1.png?width=902&height=215&name=image%20(56)-1.png "Diliforce Growth Governance Platform")
AI can make mistakes. Always review the results carefully.
Several documents
Make a visual analysis of these documents.
You are acting as a visual intelligence analyst. I am giving you a set of documents (could be 2, could be 20+). Your job is not to summarize them one by one — it's to find the structure across them and make that structure visible.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
I need a tabular comparison (clean stylish rows/columns, clear criteria, concise cell summaries, documents/items as rows, comparison criteria as columns, highlight notable risks/anomalies in light pastel red and advantages/strengths in pastel green). Before building it, ask me one question at a time:
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
Review the board meeting materials and create a Board Meeting Timeline. Present chronologically: key moments per meeting, strategic evolution, major decisions and when they occurred, initiative tracker, financial and operational trends, recurring board attention areas. End with 'What a board member should know before the next meeting.'
Act as a governance auditor. Review the chosen board materials and assess governance quality across: decision-making (are decisions clearly recorded with rationale and voting?), accountability (action items assigned and followed up?), risk oversight (risks systematically tracked?), transparency (financial information complete and consistent?), compliance signals. Score each 1–5 and provide a governance health summary with recommendations. Mark each dimension with traffic-light emojis and a warning emoji for any compliance red flags. Use actual emoji characters in your response, not the words.
I am a new board director joining next week. From the last 12 months of board materials, give me: the 5 most important strategic decisions, ongoing initiatives with status, open risks, recurring concerns, and what management will likely ask the board next.
You are my CEO strategic advisor. Analyze all board materials provided — financials, KPIs, board decks, customer data, HR reports, pipeline, and any other documents — and produce a complete pre-board executive briefing.
Surface what matters most, not what's easiest to find. Prioritize surprises, inflection points, and anything that could embarrass or elevate me in the boardroom. Flag data gaps that would change your conclusions.
Use tables, emoji markers, and sharp executive language. Be opinionated. No hedging, no filler.
For the next 3 board meetings, give me: required agenda items by statute, recurring quarterly items, decisions deferred from prior meetings that need to come back, materials needing prep, and notice-deadline calendar.
Identify the top 5–7 strategic themes mentioned across all board materials. For each, show: when it first appeared, how it evolved meeting by meeting, current status (active / resolved / stalled / dropped), and whether it was accompanied by a concrete decision or remained discussion only. Highlight themes that have stalled or been dropped without resolution. Use a green-circle emoji for active, a yellow-circle emoji for stalled, a red-circle emoji for dropped, a "new" emoji for new, a check-mark emoji for resolved. Use actual emoji characters in your response, not the words.
You are my legal risk agent. Across all these documents, surface: open litigations, regulatory deadlines next 90 days, contracts with non-standard liability terms, and IP filings that need renewal. Use red, yellow and green traffic-light emojis per item, plus a warning emoji for upcoming deadlines. Use tables, emoji markers, and visuals as needed.
Look across the last 24 months of decisions in this vault. Identify decision-making patterns: how long from issue raised to decision, which committee resolves what, decisions reversed within 12 months, decisions that drift without resolution. Suggest 3 governance improvements. Tag decisions: a check-mark emoji for resolved, an in-progress / refresh emoji for drifting, a red-circle emoji for reversed within 12 months. Use tables, emoji markers, and visuals as needed.
Make a visual analysis of these documents.
You are acting as a visual intelligence analyst. I am giving you a set of documents (could be 2, could be 20+). Your job is not to summarize them one by one — it's to find the structure across them and make that structure visible.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
I need a tabular comparison (clean stylish rows/columns, clear criteria, concise cell summaries, documents/items as rows, comparison criteria as columns, highlight notable risks/anomalies in light pastel red and advantages/strengths in pastel green). Before building it, ask me one question at a time:
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
Across all entity records, build a consolidated table: entity name, jurisdiction, type, registration number, current status, last filing date, next filing due, registered agent, directors of record. Use tables, emoji markers, and visuals as needed.
Across all these entities, list every upcoming statutory filing in the next 12 months. Group by jurisdiction. Flag any overdue and any with conflicting deadlines. Color-code: a red-circle emoji for overdue, a yellow-circle emoji for due within 30 days, a green-circle emoji for OK.Use tables, emoji markers, and visuals as needed.
Act as my entity management agent. Across all chosen subsidiary records, show: ownership chain, board composition per entity, dormant vs. active entities, entities with overdue filings, and entities ripe for consolidation or dissolution. Use a green-circle emoji for active and a red-circle emoji for dormant; a warning emoji for overdue filings. Use tables, emoji markers, and visuals as needed.
From all the chosen entity records, draw the corporate group structure: parent and subsidiaries with ownership %, jurisdictions, and active/dormant status. Use a green-circle emoji for active and a red-circle emoji for dormant. Highlight any entity where the ownership chain is ambiguous with a warning emoji. Use actual emoji characters in your response, not the words.
Across the entity records, list every person who serves as director or officer in more than one of our entities. For each: entities held, roles, appointment dates. Flag any conflict-of-interest patterns or concentration of control. Mark concentration risk with a warning emoji. Use tables, emoji markers, and visuals as needed.
Identify every entity with no material activity in the last 12 months (no board decisions, no filings beyond statutory minimum, no transactions). For each, suggest: keep dormant, reactivate, or wind down — with reasoning. Tag each entity: a green-circle emoji for keep dormant, a yellow-circle emoji for reactivate, a red-circle emoji for wind down. Use tables, emoji markers, and visuals as needed.
Across all the entity records, surface every change in shareholder composition or share class structure in the last 36 months. Output as a timeline with entity, date, what changed, and resulting ownership %. Use tables, emoji markers, and visuals as needed.
Make a visual analysis of these documents.
You are acting as a visual intelligence analyst. I am giving you a set of documents (could be 2, could be 20+). Your job is not to summarize them one by one — it's to find the structure across them and make that structure visible.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
I need a tabular comparison (clean stylish rows/columns, clear criteria, concise cell summaries, documents/items as rows, comparison criteria as columns, highlight notable risks/anomalies in light pastel red and advantages/strengths in pastel green). Before building it, ask me one question at a time:
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
Analyze all contracts and create a portfolio dashboard: summary metrics (total value, number by type, average duration), portfolio table, risk heat map by category, urgent actions for contracts expiring within 90 days or with high risk. Use tables, emoji markers, and visuals as needed.
Across all the documents, map every material obligation. Create: obligation matrix (document, party, obligation, deadline, connected obligations), conflict or overlap flags, concentration risk, recommended consolidation or renegotiation priorities.
Aggregate by counterparty: number of contracts, total committed value, earliest and latest expiry, combined liability exposure, key obligations we owe them, key obligations they owe us, overall relationship risk rating. Identify top 3 most strategically and financially significant counterparties and flag concentration risk. Use tables, emoji markers, and visuals if needed.
You are my CFO agent. From all these documents and contracts, give me: cash commitments next 90 days, late or at-risk receivables, contract renewals with financial impact >€10k, and covenant headroom. Use red, yellow and green traffic-light emojis on each item by financial risk and a warning emoji for what needs action this week. Use tables, emoji markers, and visuals as needed.
Scan everything and flag anomalies: missing signatures, expired-but-active contracts, conflicting clauses, decisions without follow-up, KPIs contradicting between reports, suddenly absent meeting attendees, contracts amended without board approval. Tag each anomaly by severity: a red-circle emoji for critical, a yellow-circle emoji for medium, a green-circle emoji for minor. Use tables, emoji markers, and visuals as needed.
You are a corporate governance analyst specializing in parent companies, subsidiaries, and ownership chains. Analyze the provided company records, documents, and metadata, and produce a McKinsey-style corporate ownership overview.
RULES: Use only the provided sources. Never invent data; mark anything missing as "Not specified." Distinguish parent companies, subsidiaries, and minority investments, and identify direct vs indirect ownership.
Produce five sections:
- OWNERSHIP MAP (main output)
A Mermaid flowchart with a top-down hierarchy: Ultimate Parent → Intermediate Holdings → Subsidiaries → Minority Investments. Mark the ultimate parent with a star (★) and show ownership % on each relationship. Each node uses two-line labels with: Company Name, Managing Director/CEO, Country. Color-code by given color on Key Information page. - COMPANY GOVERNANCE TABLE
Columns: Company Name | Role (Parent/Subsidiary/Associate) | Parent Company | Ownership % | Managing Director/CEO | Board Chair | Country | Sector | Key Governance Notes. - OWNERSHIP RELATIONSHIP TABLE
Columns: Parent Company | Subsidiary/Investment | Ownership % | Relationship Type | Direct/Indirect | Country. Types: Subsidiary (>50%), Joint Venture (50%), Associate/Minority (<50%). - GOVERNANCE ANALYSIS
5-8 concise bullet points covering the ultimate parent, main subsidiaries, minority investments, countries of operation, intermediate holdings, and potential governance or compliance risks. - EXECUTIVE SUMMARY
One sentence summarizing the group structure (ultimate parent, key majority-owned subsidiaries, minority investments, and countries).
Produce five sections:
- OWNERSHIP MAP (main output)
A Mermaid flowchart with a top-down hierarchy: Ultimate Parent → Intermediate Holdings → Subsidiaries → Minority Investments. Mark the ultimate parent with a star (★) and show ownership % on each relationship. Each node uses two-line labels with: Company Name, Managing Director/CEO, Country. Color-code by given color on Key Information page. - COMPANY GOVERNANCE TABLE
Columns: Company Name | Role (Parent/Subsidiary/Associate) | Parent Company | Ownership % | Managing Director/CEO | Board Chair | Country | Sector | Key Governance Notes. - OWNERSHIP RELATIONSHIP TABLE
Columns: Parent Company | Subsidiary/Investment | Ownership % | Relationship Type | Direct/Indirect | Country. Types: Subsidiary (>50%), Joint Venture (50%), Associate/Minority (<50%). - GOVERNANCE ANALYSIS
5-8 concise bullet points covering the ultimate parent, main subsidiaries, minority investments, countries of operation, intermediate holdings, and potential governance or compliance risks. - EXECUTIVE SUMMARY
One sentence summarizing the group structure (ultimate parent, key majority-owned subsidiaries, minority investments, and countries).
Two documents
I need a tabular comparison (clean stylish rows/columns, clear criteria, concise cell summaries, documents/items as rows, comparison criteria as columns, highlight notable risks/anomalies in light pastel red and advantages/strengths in pastel green). Before building it, ask me one question at a time:
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
You are acting as a visual intelligence analyst. I am giving you a set of documents (could be 2, could be 20+). Your job is not to summarize them one by one — it's to find the structure across them and make that structure visible.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Compare these two CEO reports. Produce: (1) theme comparison table — finance, strategy, people, risks, customers, operations × Report A | Report B | Change (improved / declined / unchanged / new); (2) commitments tracker — promises from Report A vs. status in Report B; (3) KPI side-by-side; (4) new topics and omitted topics; (5) board-level summary in 3–5 bullets.
Compare these two sets of board minutes. Identify: recurring themes and how they evolved, decisions made at each meeting (new vs. revisited), action items from meeting A and their status in meeting B, new risks in meeting B, topics present in one meeting but absent in the other. Use a "new" emoji for new items, a check-mark emoji for resolved, a red-circle emoji for new risks, a right-arrow emoji for carried over. Use actual emoji characters in your response, not the words.
I need a tabular comparison (clean stylish rows/columns, clear criteria, concise cell summaries, documents/items as rows, comparison criteria as columns, highlight notable risks/anomalies in light pastel red and advantages/strengths in pastel green). Before building it, ask me one question at a time:
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
You are acting as a visual intelligence analyst. I am giving you a set of documents (could be 2, could be 20+). Your job is not to summarize them one by one — it's to find the structure across them and make that structure visible.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Compare these two versions of the entity record. Identify changes in ownership, directors, registered office, statutory capital, or activity codes. Flag any change that may have required regulatory notification.
Compare the corporate filing requirements implied by these two entity records (different jurisdictions). Show side-by-side: required annual filings, deadlines, fees, and key local quirks.
Compare these two group-structure snapshots. Identify: entities added, dissolved, or moved; ownership chain changes; new intercompany relationships; tax-residence shifts. Output a structured diff with risk flags.
Compare these two foundational documents for the same entity. Identify: conflicts between the two, gaps where one is silent, terms where SHA is more restrictive than AoA, and provisions that may not be enforceable as drafted.
I need a tabular comparison (clean stylish rows/columns, clear criteria, concise cell summaries, documents/items as rows, comparison criteria as columns, highlight notable risks/anomalies in light pastel red and advantages/strengths in pastel green). Before building it, ask me one question at a time:
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
1. What am I comparing, and for what purpose? (e.g., contract clauses, meeting materials/board packs, legal entities, vendor options, case law, deal terms, etc.) 2. Do I have specific items and/or criteria in mind, or should you propose a sensible default set based on the purpose? When asking the clarifying questions, present numbered options I can pick from (e.g., 1, 2, 3), but also let me freely type my own answer instead. Once I answer, build the table with my specified items/criteria (filling gaps with sensible suggestions, clearly marked as suggested), keep cells short and scannable, and flag any notable risks, gaps, inconsistencies, or outliers at the end.
You are acting as a visual intelligence analyst. I am giving you a set of documents (could be 2, could be 20+). Your job is not to summarize them one by one — it's to find the structure across them and make that structure visible.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Phase 1 — Triage (do this silently, then report in one short paragraph) Identify what each document is (report, contract, transcript, dataset, email thread, etc.) and its date/version if available. Note document count and decide your synthesis strategy: 2–4 documents → deep pairwise/triangulated comparison. 5–12 documents → thematic clustering with a cluster-level diagram, then drill into notable outliers. 13+ documents → build a taxonomy first (categories/themes), quantify how many documents fall into each, then diagram the taxonomy and only spotlight the 3–5 most decision-relevant threads in detail. Flag anything unreadable, contradictory on its face, or missing expected sections. Phase 2 — Extraction For each document, extract into a working table (don't show me the raw table, use it internally): doc_id | date | key entities | key claims/figures | stated purpose | tone/stance | notable risks or gaps Phase 3 — Cross-Document Synthesis Actively look for: Agreement — claims/figures that corroborate each other across sources. Contradiction — direct conflicts (numbers, dates, positions) — name the documents involved. Evolution — how a position, number, or plan changed over time/versions. Gaps — things one document assumes but none actually establishes. Causality/dependency — where one document's output is another's input. Phase 4 — Visual Output This is the deliverable. Default to rich markdown + Mermaid.js, chosen deliberately per insight — never decorative. For every diagram, add one sentence above it explaining why this shape of chart is the right one for this insight. Use this decision guide: What you foundDiagramChronological changes across doc versions/eventstimeline or gitGraph (as a proxy for version lineage)Process, workflow, or decision logic described in the docsflowchart TD/LRWho-said-what-to-whom, or system interactionssequenceDiagramCategorization / taxonomy of themes across many docsmindmapEntities and their relationships (people, orgs, systems, clauses)erDiagram or flowchart with labeled edgesPositioning two variables against each other (e.g. risk vs. impact, cost vs. confidence)quadrantChartMulti-stage funnel or flow of volume/resources between categoriessankey-betaProject phases, milestones, deadlines mentioned in docsganttState changes (e.g. contract status, approval stages)stateDiagram-v2Simple proportions (budget split, doc types, sentiment mix)pieStep-by-step user/customer journey described in source materialjourney Additional formatting toolkit — use generously, not gratuitously: Callout blocks using blockquotes with bold labels: > **⚠️ Contradiction:** ... / > **✅ Corroborated:** ... / > **🔍 Gap:** ... Comparison tables for anything with 2+ documents sharing a dimension (dates, figures, terms, obligations). Collapsible detail via <details><summary> for supporting evidence so the main narrative stays scannable. Confidence tags inline, e.g. (high confidence — corroborated in 3 docs) vs (single-source, unverified). Short badges/emoji sparingly as visual anchors for severity or status — not as decoration. Phase 5 — Executive Layer (put this first in your final answer) One-paragraph synthesis — the single most important thing across all documents, stated plainly. At-a-glance dashboard — a compact table or small chart set: document count, date range, top 3 findings, top contradiction, top gap. Navigation — a short table of contents pointing to each diagram/section below, since the full analysis may be long. Phase 6 — Full Analysis Body Structured with headers, organized either: by theme (preferred when documents overlap in subject), or by document cluster (preferred when documents are heterogeneous), with diagrams embedded at the point of relevance (not all dumped at the end), each preceded by 1–2 sentences of interpretation and followed by "so what" — why it matters for a decision. Phase 7 — Close End with: Open questions the documents don't resolve. Recommended next document(s)/data that would sharpen the picture. A final relationship-map diagram (flowchart or mindmap) tying all documents together as nodes, so I can see the whole corpus at a glance. Rules Never invent figures, quotes, or dates not present in the documents — flag inference explicitly as inference. Prefer 3 well-chosen diagrams over 8 mediocre ones. If a diagram would have fewer than 3 meaningful nodes/data points, use a table or callout instead — don't force a chart. Keep prose between diagrams tight; the diagrams should carry the analytical weight.
Compare these two contracts in a structured table. Rows: key clauses (duration, value, payment terms, liability cap, IP ownership, termination, dispute resolution, governing law, confidentiality, renewal). Columns: Contract A | Contract B | Difference. Add red/yellow/green traffic-light emoji risk ratings per clause. Use actual emoji characters in your response, not the words. Conclude with a one-paragraph recommendation on which is more favorable and why.
Benchmark these two vendor contracts. Create a scorecard table with criteria: total value, payment terms favorability, liability exposure, SLA & penalties, exit flexibility, IP rights, renewal conditions. Score each 1–5. Highlight top 3 advantages of each. Recommend which to proceed with and flag any red-line clauses that need renegotiation.
Compare version A and version B. Identify and categorize every material change: added provisions, removed provisions, modified terms (and whether the change favors A or B), unchanged sections. Flag any changes that materially shift risk, rights, or obligations.
One document
You are an AI board intelligence assistant. Analyze these board meeting materials and produce a Board Intelligence Briefing with sections: Executive Snapshot, Strategic Themes, Key Board Decisions, Major Initiatives Progress, Financial & Operational Highlights, Risks & Issues, Upcoming Topics. Write for a busy director who needs to be ready in under 2 minutes.
Analyze this CEO report as a senior strategy advisor. Identify: strategic signals, financial health and what numbers imply, people & culture themes, stated risks vs. unstated risks, weak assumptions or missing information, three questions the board should ask management.
Read these meeting minutes and extract: all decisions made (motion, result, responsible party), all action items (task, owner, deadline), topics deferred or tabled (subject, reason, next expected date). Flag any action items with no assigned owner or deadline.
Generate a board resolution to approve [decision]. Use our existing resolution template as the format model. Include preamble, recital of facts, the resolution itself, voting structure, and signature block. Match the style of our last 3 resolutions. End with a checklist of what must be true before signing.
Polish this draft board agenda. Improve: time allocation realism, item sequencing (light → heavy → light), discussion vs. decision tagging, attachment references. Output the cleaned agenda plus a one-line rationale per change.
Clean these meeting minutes. Fix: tone (formal but readable), typos, ambiguous attributions, missing decisions, unclear action items. Preserve all substantive content. Show your changes with brief annotations.
Read this document and surface its hidden risks — the things a busy reader would skim past. Cover: clauses that shift cost or liability to us, automatic triggers, silent commitments, capped vs. uncapped exposures, and any terms that activate only under specific conditions. Output as a risk table with red/yellow/green traffic-light emoji severity, the clause reference, and a one-line plain-language explanation of what would actually happen. Use actual emoji characters in your response, not the words.
Give me a one-page reference card for this document: type, parties, effective dates, value, governing law, key obligations on each party, renewal/termination triggers, and the 3 things I would need to remember if asked about it 6 months from now. Format as a printable single page.
Decode this entity record for a non-lawyer. Show: who owns it, who controls it, what it can legally do, what filings it owes, and what would trigger a notification requirement.
Extract every regulatory or statutory deadline implied by this document. Output a table: deadline, what is required, consequence of missing it.
Read these articles of association and extract: share classes and rights attached to each, voting thresholds for different decisions, restrictions on share transfers, pre-emption rights, board powers and limitations, amendment procedure. Plain-language summary at the end.
Read this shareholder agreement and produce a compliance checklist: for each material obligation (notifications, reserved matters, drag/tag, ROFR, transfer restrictions), what evidence would prove we are in compliance and what would breach.
Read this proposed entity change (merger, demerger, share issue, capital reduction, dissolution). Identify: required corporate approvals, regulatory notifications triggered, tax implications to investigate, employee/contract impact, timeline. Output as an action checklist.
Analyze this contract and produce a one-page executive summary for a C-level reader. Include: parties & purpose, key commercial terms (value, payment, duration), top 3 obligations per party, renewal/termination triggers, and a red/yellow/green traffic-light emoji risk rating for liability, IP, confidentiality, and dispute resolution. Use actual emoji characters in your response, not the words.
Review this contract and create a risk analysis table. Columns: Clause | Risk type | Description | Risk level | Recommendation. Then add an emoji-encoded overall risk score bar for: legal risk, financial risk, operational risk, renewal urgency — use red, yellow and green circle emojis to represent high / medium / low. Use actual emoji characters in your response, not the words.
Extract all time-sensitive information. Create two tables: Key dates (event, date, party responsible, consequence of missing) and Ongoing obligations (obligation, responsible party, frequency, penalty). Flag items expiring within 90 days.
Translate this contract to [target language] with legal precision. Preserve clause numbering. After the translation, give me a one-page executive summary in the same language and flag the 3 clauses I should ask my lawyer about.
Draft a mutual NDA using our existing template. Adapt for: counterparty [name], purpose [describe], governing law [jurisdiction], term [duration], with carve-outs for [list]. Flag any clause where I should consult legal before signing.
Draft a partnership agreement first cut for partnership with [counterparty]. Base on our existing partner template. Customize: revenue split [terms], exclusivity [yes/no/territory], term [duration], exit triggers. Mark every section needing legal review.
Read this document and surface its hidden risks — clauses that shift cost or liability, automatic triggers, silent commitments, capped vs. uncapped exposures. Output as a risk table with red/yellow/green traffic-light emoji severity and a plain-language explanation. Use actual emoji characters in your response, not the words.
Give me a one-page reference card for this document: type, parties, effective dates, value, governing law, key obligations on each party, renewal/termination triggers, and the 3 things I would need to remember if asked about it 6 months from now.
Create an official Meeting Minutes Extract from the attached meeting minutes. Read the attached document and use it as the only source.
Do not invent or modify any information. First, ask the user which item(s) to include. The extract must include, where available:
- (1) Organization name
- (2) Meeting type, date, time, and location
- (3) Attendees and quorum
- (4) The selected agenda item(s) with the original decision(s) and voting result(s)
- (5) Signature blocks
- (6) The certification statement: "Certified as a true extract of the approved meeting minutes."
Adapt the terminology and formatting to match the source document (e.g., Finnish pöytäkirjanote or U.S. Meeting Minutes Extract). Produce a formal, print-ready document using the formatting style of the original minutes.
Do not invent or modify any information. First, ask the user which item(s) to include. The extract must include, where available:
- (1) Organization name
- (2) Meeting type, date, time, and location
- (3) Attendees and quorum
- (4) The selected agenda item(s) with the original decision(s) and voting result(s)
- (5) Signature blocks
- (6) The certification statement: "Certified as a true extract of the approved meeting minutes."
Adapt the terminology and formatting to match the source document (e.g., Finnish pöytäkirjanote or U.S. Meeting Minutes Extract). Produce a formal, print-ready document using the formatting style of the original minutes.
General questions
I am preparing to chair my first annual general meeting. Walk me through typical legal requirements, a standard agenda, the chair’s responsibilities, shareholder notice timelines, and the three most common mistakes first-time chairs make. Assume [jurisdiction] as the operating context. Use a table for the responsibilities and a numbered list for the mistakes. Use colored title row background.
Create a table summarizing typical board duties, with columns for area, key responsibilities, risks, and common board questions. Use colored title row background.
Provide a template for a Board meeting agenda. Include standard sections (call to order, approval of prior minutes, financial review, strategic items, risk update, AOB) and a placeholder for governance-specific items.
How to balance CEO and Board powers effectively? Outline the decision rights typically reserved for the board, those delegated to the CEO, and the gray zones where misalignment most often occurs. Include 3 practical safeguards.
Key elements in an effective ESG strategy? List the building blocks (materiality assessment, governance, metrics, targets, reporting), what good vs. weak looks like for each, and the most common pitfalls boards encounter.
Steps for conducting an internal compliance audit. Cover scoping, evidence collection, interviews, findings classification, report-out, and remediation tracking. Use a numbered list with one short rationale per step.
Please translate the above to the language I'll provide in my next comment.
Please provide a template for a risk management plan. Cover risk identification, scoring (likelihood × impact), owner assignment, mitigation actions, monitoring cadence, and board reporting. Present the framework as a table with colored title row background.
PWhat are the best practices for preparing and distributing board meeting minutes? Cover what to record vs. omit, naming voters and dissents, attaching materials, circulation timeline, approval at the next meeting, and retention. Add a checklist at the end.
What are the key compliance considerations to keep in mind when managing multiple legal entities across different jurisdictions? Show as a comparison table with columns for jurisdiction-typical areas, common pitfalls, and mitigation. Use colored title row background.
What are the best practices for maintaining legal entity compliance? Cover annual filings, director registry updates, beneficial-ownership reporting, board minutes retention, and cross-jurisdictional change tracking.
What strategies can improve efficiency in routine legal entity maintenance tasks? Show a table with several columns (task, frequency, owner, automation potential, risk of error), using font formatting and a colored title row background.
How can we identify potential compliance risks in legal entity management? Cover red flags by category (filings, signatories, beneficial ownership, related-party transactions, dormant entities), and propose 3 detection methods.
Please translate the above to the language I'll provide in my next comment.
What are the key considerations for managing an inventory of legal entities? List required fields (jurisdiction, type, status, directors, registered office, key dates), update cadence, ownership, and integration touchpoints.
What are some best practices for creating an optimal corporate structure? Cover tax efficiency, governance simplicity, liability ring-fencing, capital movement flexibility, and exit-readiness considerations.
How can we streamline the process of tracking and updating corporate records for our legal entities? Suggest a workflow, ownership model, change-event triggers (e.g. director appointment, address change), and how to keep statutory copies in sync.
Key elements of a strong contract template? Show as a table. Use colored title row background.
Draft a contract clause for a purpose I'll provide in my next comment. Use plain English, flag terms that need lawyer review, and include one or two alternative wordings.
Interpret the following legal provision I'll provide in my next comment. Explain in plain English, identify ambiguities, list practical implications, and flag what a lawyer should review.
Please provide me an NDA template in the language I'll provide in my next comment. Cover definition of confidential information, exceptions, term, return/destruction, remedies, and governing law.
Best practices for efficient contract management? Cover intake, standard clauses library, approval workflow, e-signature, post-signature obligations tracking, renewal alerts, and archival.
Strategies for negotiating favorable terms? Cover preparation, anchoring, common concession patterns (e.g. liability caps, IP, termination, indemnities), and how to read counterparty signals.
Please translate the above to the language I'll provide in my next comment.
What key clauses should be included in the contract type I'll provide in my next message? Group as: must-have, conditional, and 'nice to have'. Note common defaults that boilerplate often gets wrong.