A small finance desk, a handful of people raising and placing capital, runs five AI officers on Kavanah: a CEO, a CFO, a COO, a CRO and a CMO. Its managing partner, Henry Augustine of Baird Augustine, described it this way: "Kavanah gave us a C-suite of AI officers that actually show up every morning. By the time I've had my coffee, the pipeline, the stale deals and the three decisions only I can make are already waiting for me." This post walks through the shape of that day. It describes the setup rather than results, because the point is the pattern, and the pattern works for any small team that runs on a pipeline.
1. Morning: The Operating Check
Before anyone is at a desk, a schedule bound to the AI COO reads open tasks, overdue work and what the team is waiting on, and writes one brief: what moved, what is stuck and who owns it, and the short list of decisions only a person can make. It is emailed to the managing partner. Alongside it, a daily brief measures the day against the workspace's North Star metric, the one number the firm has said matters most.
2. Pipeline and Forecast
The AI CRO watches the Deal Room and sends a pipeline-velocity alert when deals sit in a stage too long. The AI CFO sends a revenue forecast email with charts built from the deals' stages and amounts. Each sees only what its role allows: the CRO the clients and deals tools, the CFO the finance tools. Outbound mail to investors waits for approval wherever the send policy says so.
3. The Daily Executive Meeting
The five officers meet as a boardroom. Each gives an update using read-only tools, so the numbers come from the system of record; they discuss without tools for a bounded number of rounds; and the chair closes with action items. Those become tasks with owners: a person where the item is a person's, an AI employee where it is the employee's job, and nobody at all when it is unclear which person owns it, rather than handing a human commitment to an unattended run.
4. Routing and the End of the Day
A routing schedule assigns incoming tasks to owners within the team's limits. Tasks assigned to an AI officer are worked by it, borrowing the access of the person who assigned them, never more. At the end of the day a wrap reports what closed and what slipped. Anything an officer needs decided goes on that task's Approvals section, one open ask per task and at most three per officer, and the partner sees them in one digest the next morning.
5. The Rules That Make It Work
Three, and they were learned the hard way. An AI officer never makes, approves or records a decision reserved for a person; it recommends. Every feature that runs on its own, from working assigned tasks to the daily meeting, is off until an admin turns it on. And unattended work is metered: every run is billed in credits, and an admin can cap the day's unattended spend. Deloitte reports that 48% of CFOs in a leading strategy role have already deployed AI agents in finance; Gartner expects over 40% of agentic AI projects to be canceled by 2027. The difference between the two outcomes is mostly rules like these.
Five AI officers do not replace a partner. They make sure that when the partner sits down, the pipeline, the stale deals and the decisions are already on the desk. For how to set up AI employees, and how Kavanah compares with other vendors, see our AI employees page.
