Retargeting audience from CRM segments → CSVs

The problem this solves

The most valuable advertising audience any company owns is sitting in its CRM: people who already raised a hand, at known stages of the journey. And at most companies that audience never reaches the ad platforms in usable form, because the export-transform-upload chore falls between teams. The CRM owner does not run ads; the ads team does not want to fight CRM exports; and each platform wants its customer file in its own exact format.

So retargeting runs on pixels alone - anonymous site visitors, capped by what browsers still allow - while the named audience of stalled deals, aging leads, and dormant customers goes unretargeted. Worse, when the CSV finally does get built by hand, it is a one-off snapshot: unsegmented, so the reactivation message hits current customers, and stale within weeks as the CRM moves on. The gap between owned data and paid media stays open at exactly the highest-intent end of the funnel.

How the mission runs

1. Segment HubSpot by lifecycle stage. The mission reads your HubSpot contacts and segments them by lifecycle stage - leads, MQLs, opportunities, customers, and the dormant tail - so each audience can carry a message that matches where the relationship actually stands.
2. Build platform-ready CSVs. The CSV tool formats each segment twice: once to Meta's customer file schema and once to Google's customer match schema, with the identifier columns, headers, and normalization each platform requires for clean matching. The formatting failure modes are handled once, correctly.
3. Save the set to Drive. All files land organized in Google Drive - segment by segment, platform by platform - where your ads operator can pull them for upload, and where the next refresh run can version them cleanly.
4. Write the targeting note. A one-page note accompanies the files: what each segment contains, its size, the message logic it suggests - reactivation for dormant contacts, acceleration for open opportunities, exclusion lists for current customers - and which file uploads where.

The prompt

This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.

What comes back

A structured set of retargeting CSVs in Google Drive: your HubSpot base segmented by lifecycle stage, each segment formatted for both Meta and Google customer uploads, plus a one-page targeting note explaining every audience, its size, and the campaign logic it supports - including who to exclude. The highest-intent audience you own becomes usable in paid media in the time it takes to review a folder.

Make it yours

• Cut by behavior instead: segment on deal stage, product interest, or engagement recency when lifecycle stages are too coarse for the message.
• Add suppression files: generate explicit exclusion lists - current customers, active deals - so acquisition campaigns stop paying to reach them.
• Refresh on schedule: rerun monthly so the audiences track the CRM, with the note reporting how each segment grew or shrank.

Frequently asked questions

Go deeper

• AI advertising operations: the complete guide
• Budget allocation by evidence: splits that trace to data
• How AI agents execute marketing work, step by step