A political campaign is a large coordination problem with a hard deadline. Hundreds of volunteers, thousands of voter conversations, a fundraising operation, a press operation, a candidate schedule, and an opposition that's also moving — all running through a small staff that's perpetually short on sleep. Most of the staff time is spent on coordination overhead, not on the activities that actually move votes. This guide is about putting the parts of campaign work that benefit from being together — voter outreach notes, volunteer coordination, donor context, the running strategy log — into one vault where an agent can read across them. The campaign's CRM and voter file still do their jobs. The vault is where the thinking and the soft context live. ## What campaign notes actually have to hold Campaigns fail at the seams. Between the canvasser's conversation at the door and the data entry that captures it. Between the donor call and the follow-up. Between the strategy meeting on Monday and the staff actually changing what they do by Wednesday. Each seam is a place where context drops, and most of them drop because the relevant material is in three different apps and one whiteboard. The four loads a campaign vault has to carry: - **Voter and constituent context** — beyond what the voter file holds, the soft notes from real conversations, the issues that came up at events, the comments from town halls. - **Volunteer coordination** — who's reliable, who's good at what, what's been promised, what's overdue. - **Donor and supporter context** — the conversations, the asks, the relationships, the running history. - **The strategy log** — what's been decided, what's been tested, what's working, what's not, what the candidate's saying this week. A vault holds all four together. Plain markdown for the text. Audio (a town hall recording, a donor call, a strategy meeting) transcribed with speaker labels. PDFs (the polling memo, the opposition research, the messaging deck) auto-converted to searchable text. Inline databases for the trackers that benefit from being tabular. ## Voter conversations and the soft data that doesn't fit the voter file The campaign's voter file is the structural record. Names, addresses, party registration, vote history, the ID question outcomes from canvasses. That's the database. It does its job. What the voter file doesn't hold is the soft context. The retiree on Maple Street whose husband just died and who isn't ready to talk about politics yet. The small business owner who's actually persuadable on one issue but not on two others. The neighborhood association president who's a connector and is worth investing in. That kind of context lives in the canvasser's head and dies when the canvasser stops volunteering. A "Voter context" sub-vault, with pages organized by precinct or ward, holds the soft data. After each canvass shift, the canvasser types three lines on the relevant page about anything notable — the conversation that took twenty minutes, the door that should be revisited next week, the issue that kept coming up. The data entry into the voter file still happens; this is the layer above it. Three weeks later when a precinct captain wants to know who in their precinct has been receptive to outreach on a particular issue, the agent reads the precinct page: "In the Maple Street precinct notes from the past month, find every conversation that mentioned the local school bond. Pull the relevant exchanges and the doors where the conversation was substantive." The data isn't a substitute for the voter file. It's a complement. The voter file says who voted; the precinct notes say what they cared about and how they wanted to be talked to. ## Volunteer coordination and the pages that don't exist in the CRM Volunteers are the campaign. The good campaigns treat them as long-term relationships; the struggling campaigns burn them out and replace them. The difference is mostly in the soft tracking. A "Volunteers" sub-vault, with one page per active volunteer, holds the running context. What they're good at (data entry, canvassing, phone banking, hosting events). What they've committed to. What they've actually done. The personal context — they're a teacher, they have small kids, they only have weekends. The notes from when something went wrong (the canvass shift they couldn't make because of a family emergency, handled gracefully). Before a major weekend, the field director asks the agent: "Pull every volunteer who's done at least three canvass shifts in the past month and is available weekend mornings. Flag anyone who hasn't been asked to lead a team yet but might be ready." The answer comes back in seconds, grounded in the actual record, not in the field director's memory. For volunteer recognition — the part most campaigns get wrong by either ignoring it or doing it generically — the running notes make it specific. A handwritten card that mentions the actual conversation the volunteer had at the doors three weeks ago lands differently than a generic thank-you. The same shape applies to anyone running relationship-heavy work over time; see [AI notes for executive assistants](/guides/field-service-ops/executive-assistants-ai-notes/) for a structurally similar setup with a different application. ## Donor context and the long arc of the relationship Campaign fundraising is mostly a long-game relationship problem. The major donor who gave at the maximum level six years ago and hasn't been called since. The small-dollar contributor who's actually a connector and could do an event. The person who said no last cycle but is worth a re-ask now because the issue landscape has shifted. A "Donors and supporters" sub-vault, with pages per major donor and a database for smaller-dollar contributors, holds the running context. For each major donor: bio, giving history, the issues they care about, the asks that have worked, the asks that haven't, the people they're connected to, the personal details (their kid's name, their recent illness, the project they're working on). Before a donor call, the candidate or finance director reads the donor's page. The agent helps with the prep: "Pull what we know about this donor. Summarize their giving history, the issues they've cared about, and the last three substantive conversations we've had with them. Suggest the framing for tomorrow's call." You walk into the call with real context, not with a generic script. For event planning, the agent reads across the donor pages: "Pull every major donor who's mentioned interest in education policy in the past six months. For each, summarize their last interaction and suggest who might attend a small dinner with the candidate next month." That's an afternoon of staff work compressed. ## The strategy log and the meetings that actually change what people do Most campaign strategy meetings are wasted because the decisions don't make it into the day-to-day work of the staff. The campaign manager runs the meeting, the team agrees on three things, and by Wednesday those three things have evaporated because they live in someone's head and the meeting notes are in a Google Doc nobody opens. A "Strategy" sub-vault, with a page per major strategic thread, holds the running record. Each meeting on a thread adds to the page. Decisions get marked. Action items get assigned. The page tells the story of how the campaign's thinking has evolved. For staff meetings, drop the recording onto the relevant page. It transcribes with speaker labels — useful when six people were in the room and you want to remember who pushed for which approach. The transcript lives next to the meeting notes. Three weeks later when a staffer asks "wait, did we decide to go negative on the housing issue or to stay positive?", you ask the agent: "In the past month of strategy meetings, find every discussion of the housing issue. Pull the moments where decisions got made and who was in the room." You get a real answer, with quotes, in seconds. The same commitment-capture pattern is the spine of [how to capture action items so they actually get done](/guides/meetings-people/action-items-actually-get-done/). ## Town halls, candidate forums, and the public record A campaign also generates a public record of its own. The candidate's town hall remarks. The forum debate transcript. The interview the candidate gave to the local paper. The position paper the policy team released. Drop the recordings and PDFs onto a "Public record" sub-vault. Audio transcribes with speaker labels. PDFs auto-convert to markdown. The candidate's actual words, on every issue, become searchable text. Now when the opposition mischaracterizes what the candidate said three months ago, the response is fast: "Find every time our candidate has spoken about education funding in the past six months. Pull the relevant exchanges with date and context." You get the actual quotes back, not a paraphrase, ready to push back with. The discipline of pulling from the actual record beats the discipline of pulling from press clips. For internal use — the next debate prep, the next town hall — the public record is also where the candidate's evolving position lives. Before each event, the candidate's prep doc references what they've said before so the public position stays consistent. ## Press, comms, and the messaging discipline The press operation generates its own tangle. Reporter inquiries, statements drafted, talking points distributed, op-eds placed. Most of it lives in email threads that nobody can find a week later. A "Press" sub-vault holds the running record. Pages per major reporter the campaign deals with — what beats they cover, their past coverage of the campaign, the interactions to date. A "Statements" page with every statement the campaign has issued, dated. A "Talking points" page with the current week's talking points and the running history of what's been deprecated. When a new reporter inquiry comes in, the press secretary asks the agent: "Pull what we've said about this issue in the past month. Pull the current talking points. Draft a statement that's consistent with both, focused on the angle the reporter asked about." The draft compresses the assembly part of the work; the press secretary edits and approves. ## Compliance, ethics, and the campaign as employer A practical note: campaigns operate under a thicket of campaign finance, election law, and labor regulations that vary by jurisdiction. The vault is a working tool for tracking your own work — it's not a compliance system, and it doesn't replace legal counsel on any disclosure question. The campaign's compliance officer and counsel make those calls; the vault makes the relevant material easier to find when an audit or a question arises. The same caveat applies to the campaign's role as an employer of staff and a steward of volunteer time. Capture practices, retention practices, and access controls have to match the campaign's actual obligations. The vault makes the material easier to manage; it doesn't make the obligations go away. ## A calmer way to run a campaign that's about to lose its mind Campaigns are going to be intense. The schedule is fixed by the calendar, not by the staff's bandwidth. What you can fix is the search-and-reassembly cost between everything that's happening. Put the soft context in one vault. Let the agent handle the boring searches and the structural drafting. Save the staff hours for the work that actually moves votes. Try Docapybara free — [sign up](/accounts/signup/), pick one precinct or one strategic thread, drop in the relevant volunteer notes and the last meeting recording, and ask the agent for a one-page summary the field director could brief from tomorrow.