Creative Intelligence: creatives, copy and audiences on real profit
Creative Intelligence (also called Creative Autopilot) groups your active Meta ads by the picture or video they use, by their text, and by their audience, and judges them on real orders and contribution profit. In the default RECOMMEND mode it never changes anything in Meta.
Where: Menu → Creative Intelligence → Creative IntelligencePlan: Profit, Grow and Scale
1
Run the sync
At the bottom, choose the Creative Intelligence v2 window and Audience Intelligence window and click Sync Intelligence (1). It reads Meta and your orders; it writes nothing to Meta. Refresh whole-account inventory re-reads every active campaign, ad set and ad.
Tap the picture to see it bigger. Private details are blurred.
2
Check the counters
The top cards show active campaigns, ad sets and ads, resolved creatives and the Last complete sync. Only effectively active objects are followed.
3
Read the Creative Cluster Report
Each row (2) is one picture or video and every ad that uses it: instances, spend share, real orders, real CPA, break-even CPA, contribution profit, lifecycle, fatigue, best and worst ad, tags, opportunity, Recommended Action and Confidence. Use Previous / Next to page. The boxes under it list top assets, underfunded ones, fatiguing ones, duplicates and next creative briefs.
Tap the picture to see it bigger. Private details are blurred.
4
Read Copy Intelligence
COPY INTELLIGENCE OVERVIEW does the same for your ad texts: top copy clusters, primary text, headline and CTA performance, and the creative × copy matrix.
5
Read Audience Intelligence
AUDIENCE INTELLIGENCE OVERVIEW groups ad sets by targeting: active audiences, saturation, order quality, age and gender (Meta numbers only), and audience × creative, copy and product pairs.
6
Filter the creatives table
In Filters pick a campaign, ad set, state, association, product or country, and a Sort (Net contribution, Profit ROAS, Real CPA, Spend, Evidence, Opportunity, Fatigue). Click Apply. The Effectively active creatives table shows each ad with its full economics and recommendation.
7
Set your rules
In Monitoring settings set your Target CPA, Minimum Profit ROAS, Minimum net profit, Minimum purchase evidence and the other limits, then Save settings. Keep Automation mode on RECOMMEND.
Tap the picture to see it bigger. Private details are blurred.
What each number means
Real orders: exact / pending
Exact = orders carrying the ad’s ID. Pending = attribution not settled yet. They are never mixed.
Real CPA / Break-even CPA
Spend ÷ exact real orders, next to the most you can pay per order without losing money.
Contribution profit / Net
Order revenue minus product cost and shipping (contribution), then minus ad spend (net). Unsafe means costs are missing or a currency is unclear.
Spend share
This creative’s part of the account’s spend.
Lifecycle / Fatigue state
Where the creative is in its life (new, growing, mature, declining), and whether results are dropping over several days.
Confidence
How much evidence stands behind a recommendation. Low confidence = wait for more orders.
Shadow recommendations and action audit
Every recommendation the engine made, with the evidence and the result. In RECOMMEND mode the result is always “recommendation only”.
Questions
Can it pause or scale my ads by itself?
Only if the store’s Automation mode is set to AUTOPILOT and a protected server switch is turned on by CapiTracker. By default nothing is written to Meta.
Why does it say “Insufficient exact profit evidence”?
It needs several exact real orders with known product costs. Link campaigns to products and wait for more orders.
Changing Automation mode away from RECOMMEND is for advanced users. Read every setting first, and keep the limits (maximum scale %, actions per day, maximum daily budget) low.