AI copilot

Asking questions about your deal

The copilot works inside the active workspace or project context. It is not a detached chatbot — it has access to the evidence, findings, entities, and structured data in your project.

This context matters. The quality of an answer improves when the surrounding evidence and visible files are already defined. Ask questions from within a project, and the copilot can reason across the uploaded documents, extracted data, and screening results.

The most valuable use is not “tell me everything.” It is asking the next question the team actually has in front of it: what changed, what is missing, where is risk concentrated, and what needs verification.

Use the copilot for:

• Summarising the review context — “What are the key risks identified so far?”
• Reasoning across evidence — “Which contracts have change-of-control provisions that require consent?”
• Drafting report sections — “Draft a summary of the financial findings for the phase 1 report.”
• Accelerating review — “What entities are flagged for sanctions risk and also appear as counterparties in material contracts?”

Caution

The copilot accelerates review but does not replace manual verification. Treat its output as a starting point and verify critical conclusions against the source evidence.

Ask the next diligence question, not a generic one

Real deal example: review queue triage

Instead of asking for a vague summary of the whole deal, ask “Which open requests still block the customer concentration review?” or “Which flagged entities also appear in material contracts?” Those questions line up with actual work the team needs to do next.

Data sources

Beyond the deal evidence in your project, the copilot can access external data sources for research:

Financial data:

• SEC filings (10-K, 10-Q, 8-K)
• Financial statements and key metrics
• Company profiles
• Comparable companies analysis
• M&A transaction history
• Executive compensation
• Earnings call transcripts
• Treasury rates and economic indicators
• DCF valuation inputs
• Forex rates
• Institutional ownership
• Market risk premium
• ESG data and benchmarks

Legal and compliance:

• Sanctions sweep
• Litigation scan
• Corporate registries (OpenCorporates)
• VAT registration lookup
• Patent and trademark search
• Open source licence checks
• WHOIS and certificate transparency

General research:

• Web search
• News search
• Industry classification (NAICS)

Think of these as supporting context, not the main source of truth for the project. The strongest answers combine external context with the live evidence already in Colabra.

Quick research during review

The most practical use of the copilot is fast verification without leaving the review surface.

When you are reviewing a contract and want to check whether a counterparty has litigation history, ask the copilot instead of opening a separate research tool. When you are looking at a cap table and want comparable transaction multiples, the copilot can pull them from financial data sources.

Tip

This keeps you in context. The alternative — switching to a browser, running a search, finding the right source, and switching back — breaks the review flow and adds friction that compounds across hundreds of documents.

Prompt pattern that works

• Ask about a specific review question, not the whole deal.
• Name the risk, workstream, or evidence family you care about.
• Ask for the source-backed answer or the shortlist worth checking next.
• Verify the critical items in the linked source before treating the answer as final.