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 marketing and communications

Marketing Case Study Evidence Agent

Case studies that are persuasive and fully supported.

Marketing Case Study Evidence Agent: what goes in, what the agent does and what you get

What it does

Case studies often mix approved results with loose anecdotes, which can embarrass both the company and the customer. This agent traces each outcome in the draft to customer-approved evidence, such as a signed results sheet or interview notes. It checks the strongest claims first. For numbers with weak support, it drafts clarification questions for the customer contact. Numbers that stay unsupported are removed or turned into attributed quotes. It then drafts a qualified version and checks it against the customer's attribution permissions. The customer and marketing both approve before publishing. Edge case: 'cut processing time by 50%' has no timing study, so it becomes a quote about the work feeling faster.

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 Case study draft ready 2 USES A TOOL Trace each outcome to its source 3 DOES Rank claims by strength and check the strongestfirst 4 CHECKS THE RESULT Is each number supported by approved evidence? If not: ask the customer, then remove it or turn it intoan attributed quote. Back to step 2. 5 DOES Draft the qualified case study 6 CHECKS THE RESULT Does the draft respect the customer's attributionpermissions? If not: remove names or logos not permitted. Back tostep 5. 7 YOU APPROVE Customer and marketing approve 8 RESULT Permission-aware case study
Read the steps as a list
  1. Case study draft ready
  2. Trace each outcome to its source
  3. Rank claims by strength and check the strongest first
  4. Is each number supported by approved evidence?If not: ask the customer, then remove it or turn it into an attributed quote. Back to step 2.
  5. Draft the qualified case study
  6. Does the draft respect the customer's attribution permissions?If not: remove names or logos not permitted. Back to step 5.
  7. Customer and marketing approveThe agent waits here for your OK.
  8. Permission-aware case study

How it decides

It checks the strongest claims first and downgrades unsupported ones.

  • Every number needs approved evidence.
  • Unsupported results become attributed observations, not facts.
  • Use names and logos only with written permission.

Make it yours

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

  • What counts as approved evidence
  • Attribution permissions per customer (name, logo, quotes)
  • How many clarification rounds before removing a claim (default 1)
  • Approvers on both sides

What keeps you in control

It always asks you first

  • Customer attribution
  • Publication

Hard limits

  • No unsupported numbers.

It stops when

  • Done: draft approved.

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 case study for a regional logistics client is drafted on 2 July. The agent traces 6 outcomes. The evidence check fails on 'cut processing time by 50%', which has no timing data. It asks the contact, who confirms there was no study, so it becomes 'the team said processing feels much faster'. Two other numbers match the signed results sheet. Customer and marketing approve the final draft.

More agents for marketing and communications