Skip to the main content
Interim management

Guide: finding mandates

Finding mandates with AI, without making anything up

Steffen Herr developed and tested this five-prompt routine. It runs in ChatGPT or Claude. The AI reads the CV and LinkedIn profile and sharpens the documents for applicant tracking systems. It drafts every message; sending stays manual.

What the routine is for

  • Plan the next mandate early

    Running interim mandates last 10.3 months on average (DDIM flash survey, August 2026). The search for the next one starts while the current one is still running. Time for it is short.

  • Many platforms, new requests every day

    Project boards, job alerts and agencies send suggestions daily. The work lies in sorting: which request fits the role, the length and the days per week?

  • What applicant tracking systems filter out

    Agencies and companies filter first. A CV that uses the terms of the brief gets found. The same content in other words slips through.

  • Evidence instead of embellishment

    Every prompt forbids invented projects, clients and figures. Anything not in the CV is flagged as a gap. That protects you in the conversation with the business side.

Five steps from profile to message

  1. 01 · Step 1

    Analyse the profile

    Upload the CV and the LinkedIn profile as PDF. Prompt: "Act as a senior recruiter. Name the 20 roles I am best qualified for, the skills I lack and the key ATS terms for each role. Do not invent anything."

  2. 02 · Step 2

    Build a master CV

    Prompt: "Rewrite every bullet with the XYZ formula: accomplished X, measured by Y, by doing Z. Do not invent results. If a figure is missing, ask." Variants follow from it, for example for IT roles and for transformation mandates.

  3. 03 · Step 3

    Review recent requests

    Prompt: "Find projects and interim mandates from the last five days that fit these roles. Table with link, fit, matched skills and the best CV variant." Only hits with a checked link and date count.

  4. 04 · Step 4

    Break down the brief

    For each hit the AI checks must-haves, pain points and contract type and decides: go, go with gaps or no go. Without a go there is no application.

  5. 05 · Step 5

    Prepare the message

    The message to the agency or hiring manager opens with the role and the central requirement, names the evidenced experience and the reason for a call. The AI stops before sending.

Rules that carry the results

  • Only facts from your own CV; the routine names gaps instead of glossing over them
  • Use the terms of the brief where they are backed by evidence, without keyword chains
  • CV in two versions: designed for e-mail, single-column for applicant portals
  • Sending stays manual: every message is read before it goes out

Frequently asked questions

Does it work with ChatGPT and with Claude?

Yes. Steps 1, 2, 4 and 5 only need the uploaded documents. Step 3 needs a model with web search, otherwise it knows no requests from the last few days.

Should the AI send applications on its own?

Steffen Herr advises against it. A wrong statement in a mandate request cannot be taken back. The routine prepares everything and stops before sending.

How should LinkedIn jobs be handled?

Safest through the job alerts by e-mail and the public job pages. LinkedIn's User Agreement prohibits automated access to the logged-in account; it can get the account restricted.

Does the routine help companies looking for an interim manager?

Read the other way round, yes. A brief that clearly states must-haves, length, days per week and contract type gets better-matched replies.

Related pages

A mandate to fill? An answer within 24 hours.

Availability, terms and fit settled up front. Directly, without detours.

Last updated: 14/10/2026