A sample operating-intelligence brief.

I. Executive read

Harborline is not short on effort, data, or software. It is short on a shared operating picture. The company has a reliable bookkeeper, a usable accounting file, a good sales memory inside two people, and a warehouse team that notices problems before the reporting system does. Those assets are not yet connected.

The visible result is drift: margin questions arrive too late, purchasing decisions live in side conversations, and leadership meetings spend too much time reconstructing last week's facts before deciding what to do next. AI can help here, but only after the information spine is made clearer.

II. Financial truth

Revenue is up 9.8% year over year through May, but gross margin has moved from 38.4% to 34.9%. The margin decline is not evenly distributed. Contractor accounts are holding, while direct retail and emergency-order work have absorbed freight and rush-purchase costs that are not being priced back into the sale.

The accounting file shows the trend. The purchasing team knows the reason. The sales team feels the customer pressure. No single rhythm currently puts those three truths in the same room before the month is over.

III. Operating rhythm

The Monday leadership meeting has the right people and the wrong artifact. It begins with updates, moves into exceptions, and ends with assignments that are not always visible the following week. This is not a discipline problem. It is a memory problem.

A better first rhythm would be a one-page Weekly Pulse: cash, margin signal, receivables risk, open purchasing decisions, and the three commitments that must move before Friday. The existing meeting can stay. The artifact should change.

IV. Decision memory

Harborline has made the same pricing decision three times in six months: absorb the exception now, revisit the policy later. The policy has not been revisited because the decision disappears once the customer issue is solved.

Start a decision log with four fields: decision, reason, owner, review date. Do not make it elaborate. The first test is whether the company can find its own thinking two weeks later.

V. AI leverage

The first useful AI workflow is not a sales chatbot or a replacement for the bookkeeper. It is a weekly operating narrator that reads the approved inputs and drafts the first version of the Weekly Pulse for human review.

Inputs: month-to-date P&L export, open AR, order-margin report, purchasing exceptions, and last week's commitments. Output: a draft meeting brief with the numbers that moved, decisions waiting, and commitments at risk. The leader still decides. The system stops making them reconstruct the company from scratch.

VI. First moves

1. Build the Weekly Pulse in its simplest form and use it for four consecutive Mondays.
2. Add the decision log and backfill only the three pricing decisions that are still active.
3. Price emergency-order freight explicitly for one customer segment before expanding the policy.
4. Prototype the AI operating narrator after the Weekly Pulse has two weeks of real use.

---

Composite sample. No client data is used.