How to write an AI hiring brief (ICAP / role brief).

I use this to turn a hiring-manager intake call into an ICAP - the Ideal Candidate and Assignment Profile that is the source of truth for a search. It is the brief everything downstream flows from: job ad, sourcing strings, scorecard, outreach. I reach for it at the start of every new role, when I have an intake call or notes and need the profile that anchors the whole search.

How I use this, and what for

I use this to turn a hiring-manager intake call into an ICAP - the Ideal Candidate and Assignment Profile that is the source of truth for a search. It is the brief everything downstream flows from: job ad, sourcing strings, scorecard, outreach. I reach for it at the start of every new role, when I have an intake call or notes and need the profile that anchors the whole search.

What the AI actually does

You hand it the role and the intake call. The AI pulls company data from the web, checks Granola, Fireflies and Airtable for existing context, and only asks you about what it could not find. Then it writes the ICAP in meritt's format and files it where the rest of the pipeline can pick it up. The manual version is reading the transcript, googling the company, and typing a template from scratch. This is one prompt with the gaps interviewed out of you.

The call intake matters here, and the two tools work differently. Granola captures audio straight from your device with no bot joining the call, then turns your rough notes into a structured summary after it ends. Fireflies joins as a notetaker, transcribes in 60+ languages, and pushes summaries and action items into your CRM. Either way, the AI reads the resulting record so you do not have to.

The tools and APIs you need

• Granola
• Fireflies
• Airtable
• Slack
• Claude

The right prompts

Build the ICAP

```
/icap-creator
Create an ICAP for [Client] - [Role]. The intake call is in Granola
[or paste notes]. Research the company first, check Granola, Fireflies and
the ATS base for context, then only ask me about the gaps. Write it into
meritt-ops and DM Lauren the permalink when it is done.
```

Expand it

```
/meritt-core-icap-workflow
Take the [Client] [Role] ICAP and produce the downstream set - job ad,
sourcing strings, screening scorecard, and outreach templates.
```

How to do it

Give it the role

Client, role, and the intake call (point to the [Granola](https://www.granola.ai/) or [Fireflies](https://fireflies.ai/) meeting) or paste the notes.

Let it gather context first

- web research on the company, plus a check of Granola, Fireflies and [Airtable](https://www.airtable.com/) for anything already known.

Answer only the gaps

It interviews you on what it could not find, not the whole thing.

It writes the ICAP

in meritt's format - the source of truth.

It files and shares

Written into meritt-ops and committed; on request, Lauren gets a [Slack](https://slack.com/) DM with the permalink and a paste-ready prompt.

Expand downstream

with `meritt-core-icap-workflow` for the job ad, sourcing, scorecard and outreach.

See the difference

Let AI carry the heavy lifting; you keep the judgement and the final say.

How you'll know it's working

• A full ICAP from an intake call, not a blank template waiting on you.
• It only asked about what it genuinely could not find.
• Filed in meritt-ops and shared with Lauren in one move.
• Everything downstream (ad, sourcing, scorecard, outreach) traces back to it.

Questions people ask