Buyer Kitty operates your Meta ad account. One per connected ad account, living in that account’s media buyer chat.
Its job is the spend: getting ready creatives live, scaling what works, cutting what doesn’t — always through a plan you approve first.
The rhythm: snapshot → plan → approve
Every conversation follows the same shape:
- Snapshot. Buyer Kitty reads your full account: campaign, adset, and ad structure plus performance over 7 days, 3 days, and today — and which new creatives are ready to launch.
- Plan. It drafts a set of actions, each with explicit reasoning: why this ad should scale, why that one is fatigued. The plan renders in a drawer at the bottom of the chat.
- Approve. You approve or reject. Approved plans execute on Meta automatically, action by action, each with its own status — so a single failed action never silently sinks the rest.
Nothing executes without your approval. Buyer Kitty has no autonomous mode — every budget change, pause, launch, and duplication on your account traces back to a plan you signed off on.
Plan actions
| Action | What it does |
|---|
| Launch creative | Take a ready creative live — into an existing adset or a new one |
| Scale / trim budget | Raise or lower adset or campaign budgets |
| Pause / activate | Turn ads, adsets, or campaigns off and on |
| Duplicate | Copy a winning ad into another adset, or an adset into another campaign |
| New adset / campaign | Build fresh structure for a new test or a scaling push |
You can also drive it directly: “launch HRT-1234-1 in German”, “pause everything below 1.0 ROAS”, “what would you scale today?” — it responds with a plan.
Launching creatives
When a plan launches a creative, Buyer Kitty handles the Meta plumbing:
- Uploads the media to your ad account in all published sizes.
- Picks the right ad shape — a simple single ad, a carousel, a multi-placement ad (different aspect per placement), a multilingual ad, or a flexible ad — depending on what was published.
- Applies your account defaults (below) — pixel, conversion event, page and Instagram identity, EU DSA info.
- Routes each language to the right destination — a German ad points at your German landing page.
Launching flips the creative’s language to launched on the kanban and freezes that variation — see Creo Kitty.
Account defaults
Set once per ad account when you connect Facebook (and editable later in Integrations). Every ad Buyer Kitty launches inherits them:
| Setting | What it controls |
|---|
| Pixel | Which pixel new adsets optimize against |
| Conversion event | What Meta optimizes toward — Purchase or Start Trial. Pick the event your funnel actually fires |
| Facebook page | The page your ads run under |
| Instagram identity | The IG account shown on ads (auto-fallback if you don’t have one) |
| Excluded purchasers audience | Custom audience excluded when targeting says “exclude purchased” |
| DSA beneficiary | Required “who benefits from this ad” info for EU delivery |
The playbook
Buyer Kitty follows a media-buying playbook, so its proposals are opinionated rather than arbitrary:
- Sales objective only — every campaign optimizes toward your conversion event, not clicks or reach.
- Scale winners deliberately — budget increases in steps, watching for performance decay, rather than 10x overnight.
- Cut decisively — fatigued and underperforming ads get pause proposals before they burn budget.
- Respect your targets — your minimum ROAS, target CPA, and daily spend from setup bound every recommendation.
Disagree with a proposal? Say so — the reasoning is right there in the plan, and pushing back refines both the plan and the cat’s understanding of your risk tolerance.
Good to know
- Performance data updates hourly — see The Marketing Loop → Learn.
- Existing ads are respected. Ads you launched outside the platform show up in snapshots and can be managed by plans too; their creatives appear as shadow cards on the kanban.
- Ad codes like
HRT-1234-1-de identify exactly which creative, variation, and language an ad runs — decoder in the Metrics Glossary.