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 market research analysts

Intelligence Forecast Scorecard Agent

Each quarter the analyst knows which prediction methods earned trust and which causes repeat.

AnalysisMarketing
See how it worksGet access
Intelligence Forecast Scorecard Agent: what goes in, what the agent does and what you get

What it does

Analysts make calls such as a competitor will cut prices by spring, but rarely go back to see who was right. This agent keeps a log of each prediction with its date, method and confidence. Each quarter it looks up what actually happened using named sources and scores each call as right, partly right or wrong. It groups the misses by cause, such as bad data, a changed assumption or a surprise event, and checks whether the same cause keeps coming back. When a miss looks like a data error, it returns to the source data before blaming the method. The analyst approves the lessons before any method changes. Edge case: an outcome is unclear, so the call is marked unresolved and revisited next quarter.

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
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Quarter ends or a prediction falls due 2 USES A TOOL Load predictions due in the period 3 USES A TOOL Look up what actually happened in the named sources 4 CHECKS THE RESULT Is the outcome clear and sourced for each call? If not: search a second source; if still unclear mark asunresolved and carry it forward. Back to step 3. 5 DOES Score each call as right, partly right or wrong 6 DOES Group misses by cause 7 CHECKS THE RESULT Does any miss look like a data error rather than amethod error? If not: return to the original data, correct it andrescore. Back to step 3. 8 DOES Test whether the same cause recurs across quarters 9 YOU APPROVE Analyst approves the lessons before methods change 10 RESULT Scorecard and lessons filed
Read the steps as a list
  1. Quarter ends or a prediction falls due
  2. Load predictions due in the period
  3. Look up what actually happened in the named sources
  4. Is the outcome clear and sourced for each call?If not: search a second source; if still unclear mark as unresolved and carry it forward. Back to step 3.
  5. Score each call as right, partly right or wrong
  6. Group misses by cause
  7. Does any miss look like a data error rather than a method error?If not: return to the original data, correct it and rescore. Back to step 3.
  8. Test whether the same cause recurs across quarters
  9. Analyst approves the lessons before methods changeThe agent waits here for your OK.
  10. Scorecard and lessons filed

How it decides

It scores a call right when the outcome matches within the stated range, and flags a cause as recurring when it appears in 3 or more misses.

  • Score right only when the outcome lands inside the stated range.
  • Call a cause recurring at 3 or more misses.
  • Carry unresolved calls forward for up to 2 quarters.
  • Compare confidence level with actual hit rate for each method.

Make it yours

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

  • Scoring scale and ranges
  • Review cadence (default quarterly)
  • Outcome sources by topic
  • Number of misses that mark a cause as recurring (default 3)

What keeps you in control

It always asks you first

  • Analyst approves the lessons before any method changes

Hard limits

  • Never alter a past prediction entry
  • Show the source for every outcome

It stops when

  • Done: scorecard approved and lessons filed.
  • Stop: fewer than 5 resolved calls, so scoring waits for the next quarter.

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 happensIn Q3 the log held 12 predictions. One said a rival would launch in August. The outcome check failed because the launch was only announced for beta. The agent searched a second source, found a general release on October 1, and scored it partly right. Misses from late source data appeared 3 times, so the analyst approved a rule to confirm dates twice.

More agents for market research analysts

Decisions · Research

Market Entry Screening Agent

A scored shortlist of markets based on complete data.

10 steps2 checks11 jobs
Analysis · Reporting

Win and Loss Interview Synthesis Agent

Show the real reasons deals are won and lost, with evidence strength for each.

11 steps1 check8 jobs
Monitoring · Reporting

Competitor Earnings and Filing Review Agent

Turn each competitor filing into a short, verified alert about what changed.

10 steps1 check9 jobs
Reporting · Analysis

Competitor Hiring Signal Agent

Produce a weekly digest of competitor moves inferred from hiring, each with a confidence level and sources.

11 steps2 checks8 jobs