The job
A report is due on Friday. Someone spends three hours pulling data from the CRM, the accounting system, the project tracker. They build the spreadsheet. They format it. They write the narrative. They send it. Next Friday, they do it again. By year end, this is ninety hours of repetitive work.
The dish reads your source data on a schedule. CRM pipeline. Accounting actuals. Project status. It builds the report from a template you define. Same format. Same metrics. Same narrative structure. It lands in leadership’s inbox Friday morning. By week two, the reporting happens in the background. The human approves and sends.
Plated well, this looks like: consistent reports on a schedule. The numbers are always current. The format is always the same. Leadership can compare week to week without decoding a new layout. The narrative is there, drafted, waiting for eyes.
The recipe
All seven ingredients still apply. The leverage is Measurement (Ingredient #6) and Output Over Process (Ingredient #5). The station needs to know what done looks like. Not the steps to make the report. The output. A one-page board summary. A dashboard with six metrics. A narrative memo. You define the destination. The station figures out the steps.
Context (Ingredient #2) is the input. Where is the data? Which database. Which API. Which spreadsheet. The station needs to know where to look and what to read.
Examples (Ingredient #4) are three reports you wrote yourself, in your voice, in your format. The station learns from them. Training (Ingredient #1) is your reporting standard. Numbers are always currency. Metrics round to whole numbers except for percentages. The title format is always “Weekly Board Report: Week of [date].”
How to build it
-
Define the report structure. Weekly board report has six sections: pipeline, revenue, capacity, risks, wins, and next week. Monthly financial report has four sections: P&L summary, cash forecast, variance narrative, forward look. Write it down. This is the skeleton.
-
Define the metrics. Pipeline report: open deals by stage. Revenue report: actual versus budget by line item. Capacity report: utilization by role. These are the numbers that change each cycle.
-
Identify the source systems. Pipeline comes from the CRM. Revenue comes from accounting. Capacity comes from the project tracker. Write down the data path for each metric.
-
Write three reports by hand. Real reports from the last three cycles, in your voice, in your format. This is the training set. The station learns what “good” looks like.
-
Define the narrative rules. If actual is below budget by more than ten percent, mention it. If a metric hasn’t moved in three weeks, flag it. These rules are the guardrails. They tell the station what to call out.
-
Build the template. The report structure in the format you send: PDF, Google Slides, an email, a Slack message. The station delivers into this template.
-
Run a test cycle. Set the station to build the report on Friday at 4 PM. Have it land in your inbox. Compare it to the report you would have written. What’s missing. What’s different. What’s better.
-
Measure accuracy. Are the numbers right. Is the narrative meaningful. Do you need to change anything before sending to leadership. By week three, you’re only checking. By week four, you’re skipping the check.
What breaks it
-
The template is too loose. The report structure is unclear so the station generates different sections each week. Leadership can’t compare week to week because the layout changed.
-
Metrics definitions are fuzzy. You want “revenue” but revenue could be invoiced, collected, or accrued. The station picks one. The numbers don’t match what finance publishes. Leadership notices. Trust erodes.
-
The source data is stale. The CRM was last updated three days ago. The station pulls it Friday morning. The pipeline report is accurate but three days out of date. Leadership makes decisions on lagging data.
-
No feedback mechanism. The station generates a narrative. Leadership thinks it’s missing something important. Nobody tells the station. Next week’s report is the same. The feedback never makes it back.
When it’s working
By week one, the report drafts automatically on schedule. Ninety percent of the content is perfect. Ten percent needs human review or narrative adjustment. By week two, the human is only reading for narrative tone. Numbers are trusted. By week three, the report goes directly to leadership with only a spot-check from ops. By week four, the report is a background process. Leadership wakes up Friday morning to the brief. The station owns the work.
The signal that the recipe is sharp: a metric changes unexpectedly and the station’s narrative automatically calls it out as a risk because it learned the pattern, not just copying static language.
Monday Move
Write the report for last week by hand. Document every data source you touched, every number you checked, every sentence you wrote. Show the station. Have it write this week’s report while you watch. Where does it match your work. Where’s the gap. That gap is what needs training.
Dish 3 of 10 on the Operations Station. Build-note leverage: Measurement (Ingredient #6) + Output Over Process (Ingredient #5).