Context
aiassure helps diligence teams move from request lists to locked reports in one workspace. Industry scope packs, typed document parsers, normalization workflows, and a deal copilot grounded in live deal data keep AI under CPA control.
The problem
M&A diligence teams juggle spreadsheets, disconnected workpapers, and ad hoc AI tools that are hard to audit. Buyers need defensible EBITDA bridges, working capital pegs, and evidence-linked findings, with every adjustment approved before it affects the report.
How we approached it
We built a modular workspace where deterministic engines handle EBITDA, net working capital, DSO/DIO/DPO, and concentration math while AI assists extraction, narrative, and review. A deal copilot calls tools over live deal data, streams markdown answers with charts, and routes uploads into the same parsers as the document module. Mutations stay on an audit trail with human approval gates.
What shipped
- Multi-tenant firm workspace with org roles, pipeline, and industry templates
- Request list, document room, and typed parsers for trial balances, GL, aging, and contracts
- Normalization, QoE, working capital, concentration, and extended diligence modules
- Deal copilot with threaded chats, tool calls, inline charts, and document uploads
- Report locking with PDF, PowerPoint, and Excel export paths
Outcomes
- End-to-end diligence workflow scoped per deal from quick look to full lender packs
- AI proposals stay subordinate to CPA review before adjustments hit the bridge
- Deal copilot answers grounded in live EBITDA, issues, documents, and adjustments
- Audit trail and evidence badges support defensible review and handoff
Stack
- Next.js
- TypeScript
- Vercel AI SDK
- Multi-tenant SaaS
- Deterministic calculation engine
- Document parsing pipelines
- PostgreSQL