The deal closes and the integration plan starts looking thinner than it did during diligence. The PowerPoint that mapped the first hundred days was high-resolution at the board level and roughly nothing at the workstream level. The integration lead is now juggling fifteen workstreams, each owned by a different person, each producing its own meeting notes and its own decisions, and the only place those threads come together is in the integration lead's head.
That's where most acquisitions quietly start going sideways. Not because the strategy was wrong, but because the integration sprawl outran the working record. A vault of plain markdown notes with an integrated agent isn't a substitute for an integration management office, but it does make the connective work tractable for the integration lead who's actually running the days. The same shape underwrites due diligence in acquisitions — same documents, same agent, just on the front end of the deal.
A vault shaped like the integration plan
The shape that holds up across deal size is roughly: one top-level page for the integration, with sub-pages for each workstream — finance, people, IT, customers, product, legal, comms, ops. Each workstream sub-page holds the running brief, the decisions log, the open questions, and the meeting recaps. A separate top-level page for the playbook — the high-level milestones, the dependencies, the redlines.
Capy supports unlimited page nesting, so heavy workstreams can fan out into per-system or per-team sub-pages without forcing the structure on lighter ones. The whole vault is plain markdown. That matters because the integration lead can ask one question across every workstream — "what's slipping past its target date and why" — and get back a synthesis that would otherwise take an hour of clicking through dashboards.
Workstream owners with running briefs, not status decks
The default integration setup asks every workstream owner to produce a weekly status deck for the integration management office. The decks consume a lot of owner time, get read at a glance, and stop reflecting reality by Wednesday.
A working alternative: each workstream owner runs a one-page running brief in the vault. They drop short updates as work happens — decisions made, followups assigned, things they're stuck on. Two minutes at a time, throughout the week. By Friday, the brief is the status. The integration lead asks the agent: read every workstream's brief and surface anything that contradicts what's in the playbook, anything that's slipped without flagging, anything where two workstreams seem to be moving in different directions on the same topic.
You get a synthesis that's actually built from the workstreams' own writing, not from a decks-of-decks abstraction. (The same running-brief pattern is what makes investor-relations updates writable in a sitting.)
Decisions written with the rationale, not just the outcome
Integration decisions get made fast. The reason you collapsed the two CRM stacks into the acquirer's. The reason you kept the target's brand for twelve months. The reason you laid off the director-level redundancies in week three rather than week ten. Each one made sense in the moment. Six months later, when the question comes back from the board or from the team or from a frustrated former employee, the rationale is gone and the decision looks arbitrary.
The fix is mechanical. Each workstream page has an inline decisions database via the :::database::: directive — rows for date, decision, rationale, owner, what triggered it, and what we'd want to revisit. After every working session or steering committee meeting, ask the agent to read the recap and propose entries. You confirm or edit; the rationale lands in writing while it's still fresh.
When the question comes back, you both look at the same row. The conversation is about whether the rationale still holds, not about what people remember saying. (The same decisions-log discipline is the spine of contract negotiation with AI notes.)
The open-questions database the steering committee runs on, with diligence PDFs alongside
Integration produces hard questions faster than the answers come back. The customer overlap question that nobody's modeled yet. The product-roadmap reconciliation that's still being negotiated. The leadership decision that's waiting on the board. The default failure mode is that these questions live in someone's head, surface in the steering committee, get partial answers, and re-surface two weeks later because nobody captured what was decided.
The fix is an inline open-questions database in the integration's top-level page — rows for date raised, who raised, question, current best answer, status, owner, target close date. The database lives directly in the page via the :::database::: directive.
Before each steering committee, ask the agent to read the database and surface questions that have aged past their target close date with no update, plus any questions whose current best answer contradicts a recent decision. The committee meeting becomes a working session against an actual list, instead of a half-remembered round-robin.
The same page is where the diligence PDFs live — the target's contracts, IP filings, HR policies, third-party reports. Drop them in; they auto-convert to markdown via docstrange so the agent can read every line. Ask the agent to summarize the target's customer-contract template and flag the assignment clauses, or to read the HR handbook and find every benefit that's structurally different from the acquirer's. The questions in the database often resolve once the agent has actually read the relevant pages of the data room.
Customer and people threads that don't get lost
The two integrations that quietly hurt the deal aren't on the workstream chart. They're the customer who quietly churns because the post-acquisition pricing letter felt different from the pre-acquisition relationship, and the high-performing IC who quietly resigns because a manager said something hand-wavey about role consolidation in the all-hands.
A working setup: a "customer signal" page and a "people signal" page in the integration vault, where you drop quoted snippets from emails, slack threads, and calls as they land. Two minutes per snippet. The integration lead asks the agent weekly to read both pages and surface patterns: which customer concerns are recurring, which people questions came up multiple times, which feedback contradicts the comms plan.
You catch the slow drift instead of being surprised by the announcement. The same signal-capture habit is the spine of how account managers keep client context from slipping — different scale, same compounding observation.
Recordings with speaker labels for the harder rooms
Most integration meetings don't need to be recorded. The ones that do are the all-hands where the target's team hears the comms plan for the first time, the executive working session on org design, the steering committee where two CEOs disagree on a roadmap call. Those are the rooms where a transcript would help and where nobody wants to be the one taking notes.
Record the meeting in Capy. The transcript comes back with speaker diarization — labels like "Speaker 1: …" so you can tell who said what, not just a wall of text. Park the recording on the relevant workstream page. Ask the agent to draft a recap with three sections: what we agreed, what's still open, what we agreed to revisit and when.
The transcript matters less for the meeting itself and more for the next conversation, when somebody says "we agreed in the all-hands to handle X this way" and the room needs the exact wording.
What this isn't
Capy isn't an integration-management-office tool. The Gantt charts, the workstream RACI matrix, the board-state tracking — those still live where they live. Capy is for the unstructured side — the workstream briefs, the decision rationales, the open-questions database, the customer-and-people signal — which is the part that's currently sprawled across Slack and decks and the integration lead's memory.
It's also single-user by design. One integration lead, one vault. If your integration model is a team of fifteen people who all need to edit the same artifacts with role-based permissions, that isn't this product. The shape that fits is the integration lead — usually a chief of staff or a dedicated PMI lead — running the connective layer alongside the team's structured tools. Pricing tiers are on the pricing page.
Also: Capy doesn't and won't claim regulatory certifications relevant to specific industries. If your acquisition is in a regulated space, treat the agent as the connective layer for your team's writing — not as the system of record for compliance artifacts. Those still live in the tools your auditor expects.
A small first test
Pick the workstream you're most worried about. Drop the last two meeting recaps, the relevant section of the integration playbook, and one diligence PDF onto a Capy page. Ask the agent to draft a one-page status: what's on track, what's slipping, what's the open question that needs the most attention this week. If the agent surfaces a tension between the playbook and the recaps that you hadn't named, you've got the sense of what it does for you across the rest of the integration.
Try Docapybara free. Load one workstream's last two weeks and see what the agent finds.