Fractional work has a specific failure mode the salaried equivalent doesn't. A full-time CFO carries one company's context in their head all week; a fractional CFO is carrying five, and switching between them every two hours. The cost isn't the work itself — it's the fifteen-minute warm-up at the start of every call where you're trying to remember whether this company is the one with the EBITDA recast in process or the one with the auditor change.
A working notes setup for fractional work isn't an upgrade. It's table stakes, and the table stakes most fractional operators are running without. The fix is one vault per client where the operating record lives, and one agent that can read across each client's pages on demand so you walk into every call already in context. Adjacent multi-client work like agency owners scaling without hiring and franchise owners running multiple locations hits the same per-client structure.
A vault per client, not a folder per project
The shape that holds up across multiple concurrent engagements is one top-level page per client — Acme CFO engagement, Globex CMO engagement, Initech head-of-product engagement — with sub-pages underneath for the specific working surfaces: the operating cadence, the open issues, the board materials, the team interactions, the strategic projects in flight. Capy supports unlimited page nesting, so a deeper engagement that's been going for six months can grow per-quarter pages, per-board-meeting pages, and per-project pages without forcing you to re-architect.
Plain markdown matters here because the agent can read across one client's pages in a single query. Five minutes before the Wednesday Acme call, you ask Capy to give you a brief on the open issues, the last call's commitments, and anything that's moved in the data since the last meeting. The brief takes the time it takes to read it. You walk into the call already in context.
A pre-call brief that actually warms you up
The fifteen-minute warm-up before a fractional call is the most expensive recurring cost in the work. Multiplied across five clients and a normal week of standing meetings, that's the better part of a day spent rereading your own notes.
Inverting this: don't reread, ask. Before each standing call, ask Capy to read the client's vault and produce a one-page brief: what we discussed last time, what they committed to, what I committed to, what's changed in the data since, the three open questions worth raising on this call. The agent walks the relevant pages and gives you the brief in the time it takes to top up your coffee.
This isn't a fancier note-taking habit. It's the difference between starting the call by saying "remind me where we left off" and starting it by referencing exactly what they said three weeks ago, by name, with the right number.
Capture each client's calls with speaker labels
The classic fractional-call mistake is the four-line summary you type during the call. You're a participant and a note-taker simultaneously, and both jobs suffer. By the next week, the summary has lost the texture of who said what.
Record each call in Capy. The transcript comes back with speaker diarization — labels like Speaker 1: … so you can tell when the founder pushed back versus when the CEO conceded. Park the recording on the right call sub-page. Ask Capy to pull out:
- Decisions made on the call (and who made them)
- Action items, with owner
- Open questions for the next session
- Anything the client said that contradicts what they said in a prior session
The third bullet is the one fractional operators care about most. Founders contradict themselves all the time across calls — not maliciously, just because they're holding twenty things at once. Spotting the contradictions early is part of the value you bring.
A decision log per client that survives the year
Fractional engagements often span 12–24 months. Over that time, dozens of small decisions get made — pricing changes, hiring rationale, vendor switches, board-meeting positions. The default record is a Slack thread or a Google Doc that decays after the first quarter.
A decisions inline database per client, with rows for date, area, decision, owner, and rationale. The database lives directly in the markdown page via the :::database::: directive — alongside the prose context, not in a separate tracker. After each call, ask Capy to read the recap and propose decision-log entries for what got resolved.
When the founder asks "wait, why did we land on the 60-day notice period for that vendor," the answer six months later is row 47 in the database with the reasoning intact. The shape of the answer is "here's the thread" instead of "let me think about that."
Board prep that doesn't take a Saturday
Fractional CFOs and CMOs do a lot of board prep, and most of it is the same shape: pull the metrics, write the narrative, surface the asks, anticipate the questions. The expensive part isn't any single board pre-read; it's that each one feels like starting from scratch.
A board prep sub-page per client holds the metrics history, the narrative archive, the recurring asks, and the questions board members tend to ask. Drop the latest metrics export onto the page; ask Capy to draft the next pre-read in the structure your prior pre-reads used, with the latest data, the open issues from the decision log, and the asks worth raising. You edit, send, and move on. The full version of the board cycle is in board-of-directors meeting notes.
PDFs from auditors, lawyers, or the board itself auto-convert to markdown via docstrange so they're searchable text the agent can quote. The board package becomes a living document the agent helps you keep current, not a Saturday-night assembly job.
Cross-client research that compounds
A fractional executive ends up doing a lot of research that's reusable across clients: comp benchmarks, vendor evaluations, term-sheet norms, board-meeting cadence patterns, the right way to handle a particular kind of crisis. Done once for client A, the work should benefit client B six months later.
Park the research on a separate cross-client research vault — outside any single client's page, since the material isn't client-specific. Drop in source material: blog posts, podcast transcripts, PDFs (which become searchable via docstrange), Capy's web_search results when you need current public information, and your synthesis. The next time the same question comes up at a different engagement, the research is already there. (The portfolio-level version of this — same patterns across units — is in AI notes for franchise owners running multiple locations.)
This is the kind of compounding that solo operators rarely build because the capture cost feels too high. A vault makes it the cheaper option — not the more disciplined one.
A handoff page that protects your time
The other recurring cost in fractional work is the unscheduled "quick question" from a client between calls. Most of these are answerable without you, if the client could find the answer.
A handoff sub-page per client holds the FAQ-style answers: what we decided about pricing tiers and where the rationale lives, who owns vendor X relationship, what the cadence is for the auditor handoff. Each entry links back to the underlying decision page or research note. When a client asks the same question for the third time, you point them at the handoff page instead of writing a fresh response.
This isn't a substitute for the relationship. It's a way to keep the relationship from being defined by the same recurring question.
What this isn't
Capy isn't a fractional-executive marketplace, isn't an accounting system, and isn't a substitute for the judgment the client is paying you for. The vault holds the operating record of each engagement — the calls, the decisions, the rationale, the briefs, the cross-client research — which is the part that's currently spread across notebooks, calendar invites, and a Saturday-night sync session.
It's also single-user by design. The fractional operator owns each client's vault. Outputs ship to clients as briefs, recommendations, and board materials through the channels they already use. The shape fits the solo operator who's accountable to the client for the picture, not a multi-user team workspace.
A small first test
The cheapest way to see whether this fits your engagement portfolio is to pick one client and rebuild the last month of work in a Capy page. Drop the recordings of the standing calls, the materials you've been working with, the open issues you've been tracking. Ask Capy to draft the pre-call brief for the next standing meeting. If the brief is sharper than the warm-up you'd otherwise do — and faster to produce — you've got a sense of what running every engagement in this shape would feel like.
Try Docapybara free. Load one client's last month and see what the agent does with the engagement.