Role
Hire an AI data analyst
It settles what active user means, writes the query that matches the definition, and turns the numbers you hand it into a paragraph an executive can read.
What this worker is
An AI data analyst in Polaris is a worker you hire for the written half of analytics: defining metrics without ambiguity, specifying the query that matches the definition, interpreting result sets you provide, and producing recurring reports as documents and CSV files. It writes and reasons about analysis rather than connecting to a warehouse and running it.
- Typical output
- Metric definitions doc plus CSV
- Core connections
- Supabase, Stripe, Google Drive
- Illustrative task
- 2.5 human-hours, $5.00
What this worker is great at
Ending the argument about what a number means. Most reporting disputes are definitional: three dashboards say three different things about active users because nobody ever wrote down whether a login counts, whether internal accounts are excluded, and what the window is. This worker produces the definition document, one metric per section, with the edge cases named and the query logic spelled out beside it.
It is also good at the writing that surrounds a number. You paste in the week's figures and get back the commentary: what moved, what did not, which movement is inside normal variance and which one deserves a task. That paragraph is usually the reason a dashboard existed in the first place.
For recurring work it produces the same structure every time, which is what makes week nine comparable to week two.
What to connect it to
Supabase
The product database when Polaris-shaped teams keep their data there. A restricted, read-only key is the sensible default.
Stripe
Revenue, refunds and invoice data, which is where half of the metric definitions in a software business end up pointing.
Google Drive
Where the recurring reports have to land so the rest of the company can find them.
Web search
For benchmark definitions and how a metric is conventionally calculated, cited rather than asserted.
What a definitions task costs
Illustrative, computed from the published human-hours formula. Six metrics defined and documented in one session.
| Observed effort | Human-equivalent minutes |
|---|---|
| Picking up the task | 15 |
| 2 live web searches at 12 min each | 24 |
| 5,400 characters of finished prose at 90 chars/min | 60 |
| 4 acceptance criteria ticked at 8 min each | 32 |
| 1 progress comment, 1 doc produced | 20 |
| Total: 2.5 human-hours at $2 | $5.00 |
Where the human stays in charge
The analyst prepares
- Metric definitions with edge cases named
- Query logic that matches each definition
- Written commentary on a result set
- Recurring report structure, held steady week to week
Your data team decides
- Whether the definition is the one the company adopts
- Running anything against production
- Whether a number is trustworthy enough to act on
- What the business does about the movement
Questions people ask
+Can the AI data analyst query our database?
It writes the query and explains what it returns; running it stays outside the machine session. Supabase and Stripe are in the connection catalog for the credential side, and the analytical output is documents, commentary and files.
+How do I get useful commentary on weekly numbers?
Paste the result set into the task and state what a normal week looks like. The commentary is only as good as the context: without a baseline, any movement can be described and none of it can be judged.
+Can it build a dashboard?
It writes the specification for one: the metrics, their definitions, the cuts that matter and the queries behind each tile. Building it in your BI tool is a person's job, and the specification is the part that usually takes longest anyway.
+How does it handle a metric nobody has defined?
It proposes a definition, names the edge cases that could change the number, and marks the choices you have to make rather than quietly picking one. Those choices are what a data team should be arguing about.
Related
Hire an AI financial analyst
You supply the figures; it writes the explanation of what moved, why it matters and which assumption the whole thing rests on.
Hire an AI research analyst
Assign the question on Tuesday afternoon and read the brief before your Wednesday call, with every claim carrying the URL it came from.
Three teams, three definitions of active user, one meeting
Every definition of a metric found and listed side by side, with the source of each.
The Monday metrics email that somebody writes on Sunday night
The numbers pulled and the paragraph written, delivered as a file rather than a link.
Can you pull the numbers, sent as a direct message at 6pm
Every ask restated as a question, with the clarifying questions asked before anyone writes SQL.
Thirty requests in the backlog and one analyst
A scoping note per request, so the backlog is triaged on value rather than on arrival order.
Connect Supabase to Polaris
Supabase is the connection where a careless key choice does the most damage, and the only one Polaris cannot yet check for you.
Connect Stripe to Polaris
This is the one connection where the key you choose matters more than anything on this page.