Smart keyword grouping by intent
5 000 keywords organized into thematic ad groups in 2 minutes thanks to semantic embeddings. +15% average Quality Score.
Free audit · no credit card · 2 minutes
Account structure is the thing nobody wants to touch
And yet it's the structural decision that moves your Quality Score and CPC the most.
Bloated ad groups = capped Quality Score
20+ keywords in a single ad group make it impossible for Google to serve a relevant ad for all of them. Quality Score stalls at 5-6/10. Your CPC stays high.
Restructuring 5 000 keywords = 3 days of work
Doing it by hand in a Google Sheet takes days. Nobody gets around to it. The tech debt compounds. Performance plateaus.
Existing tools do naive pattern matching
Grouping by « contains the word X » misses intent. « running », « jogging » and « course à pied » often belong in the same ad group — pattern matching can't see that.
Under the hood: embeddings + clustering
The AI understands meaning, not just characters.
Semantic embeddings
Each keyword becomes a 1 536-dimension vector that captures its meaning. Two synonyms (« car insurance » and « auto insurance ») land close together in that space — even without sharing a word.
HDBSCAN clustering
A density-based clustering algorithm forms natural groups without you having to tell it how many to find. It also flags outliers (unique keywords that deserve their own ad group).
Visual human validation
The result is shown as an interactive map: each cluster is a bubble with its 10 most representative keywords. You can merge, split, rename, or drag a keyword between clusters.
1-click Google Ads export
The final mapping is pushed directly into your Google Ads account (new ad groups created, keywords moved, original ads kept or regenerated if you also enable Creative AI). Rollback available for 14 days.
Structural gains
Quality Score +15% on average
Better keyword↔ad relevance = better expected CTR = Quality Score up = CPC down.
CPC -20% on average
Mechanical consequence of the higher Quality Score (Ad Rank = Max bid × QS).
5 000 keywords in 2 minutes
Where a human takes 3 days, the algorithm finishes in 2 minutes.
History preserved
Keywords keep their historical Quality Score when moved (no « reset » from Google).
14-day rollback
Don't like the new structure? Restore the previous layout in 1 click for 2 weeks.
Continuous refresh
Add 100 new keywords and the AI auto-assigns them to existing clusters — or suggests new ones.
Manual vs SteerAds
Who it's built for
This feature serves very different profiles. Here's how each uses it.
20 000 SKUs organized without pain
A fashion retailer with 20k products typically has 50 000+ keywords. Manual restructure is literally impossible — ads managers live with the debt. AI clustering solves it in 30 minutes of validation.
Take over an existing account in one shot
You pick up a « legacy » client with messy structure. Instead of 2 weeks of rework, you run clustering, validate in 1h, and restart on clean ground from day 1.
Prep the move to PMax
Performance Max needs a clean asset-group structure. Pre-PMax keyword clustering gives you the basis for coherent search themes inside PMax.
Frequently asked questions
Ready to try it on your account?
Connect Google or Bing Ads in 30 seconds. See results from week one.
Start my free auditFree audit · no credit card · 2 minutes
Related features
Automatic optimization of your Google & Bing Ads campaigns
AI adjusts your campaigns' bids, budgets and statuses 24/7. On average, our users see a 17.8% CPA drop in 30 days — without raising their budget.
Forecast your campaigns at 30, 60 and 90 days
AI projects your CPA, ROAS, conversions and spend per campaign. Simulate a budget change or CPA target before committing. No more blind decisions.
Complete audit of your Google & Bing Ads account in 3 minutes
/100 score on 5 axes + 20 prioritized recommendations. Free, no credit card, no agency required.