Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for data analysts

Ad Hoc Data Request Scoping Agent

A clear request, a first result that ties to known totals, and an analyst-approved answer

Ad Hoc Data Request Scoping Agent: what goes in, what the agent does and what you get

What it does

Ad hoc data requests are often vague, and the analysis answers the wrong question. This agent reads the request and identifies the metric, time frame and definitions. It checks what data is available and asks the requester clarifying questions. It then runs a first query and tests the numbers against known totals, such as finance revenue. If they disagree, it returns to the requester with the difference and options. The analyst approves before results are shared. Edge case: the request says active customers, but the company has three definitions, so the agent asks which one and shows each count.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
ApprovedYes, continueApprovedNo 1 STARTS WHEN Data request arrives 2 USES A TOOL Read the request and look up the metric definitions 3 DOES List the metric, time frame, filters and openquestions 4 YOU APPROVE Analyst approves the questions to the requester 5 USES A TOOL Send questions and record the answers 6 USES A TOOL Run the first query 7 CHECKS THE RESULT Do the results tie to a known total within thetolerance? If not: find the difference, adjust the filters ordefinition and rerun. Back to step 3. 8 DOES Write the result with definitions and caveats 9 YOU APPROVE Analyst approves before sharing 10 RESULT Answer delivered with the query saved
Read the steps as a list
  1. Data request arrives
  2. Read the request and look up the metric definitions
  3. List the metric, time frame, filters and open questions
  4. Analyst approves the questions to the requesterThe agent waits here for your OK.
  5. Send questions and record the answers
  6. Run the first query
  7. Do the results tie to a known total within the tolerance?If not: find the difference, adjust the filters or definition and rerun. Back to step 3.
  8. Write the result with definitions and caveats
  9. Analyst approves before sharingThe agent waits here for your OK.
  10. Answer delivered with the query saved

How it decides

The request is ready when metric, time frame and definition are clear. Results are accepted when they tie to a known total within a set tolerance.

  • Ask when a metric has more than one definition
  • Tie results to a known total
  • Accept a gap up to the tolerance
  • State every definition in the answer

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Tolerance for ties (default: 1%)
  • Known totals list
  • Data sources
  • Question templates

What keeps you in control

It always asks you first

  • Analyst approves the questions and the final results

Hard limits

  • Never shares results without the analyst
  • Never guesses a definition

It stops when

  • Done: the answer is approved and delivered
  • Stop: the requester does not clarify

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA manager asked how many active customers there were last quarter. The agent found three definitions and asked which one. She chose purchase in 90 days. The first query returned 14,200, but finance reported 13,100 for the same rule. The check failed, and the agent found test accounts. After removing them, it matched at 13,080, and the analyst approved.

More agents for data analysts