What is a context hint?
OpenAI describes context hints as information attached to an ad group that helps its systems understand the conversations, needs or situations where a product or service may be relevant. A hint can add detail about what the product does, who it helps or when it is useful.
Crucially, context hints are not exact-match keywords, audience targeting rules or guarantees of delivery. This changes how they should be written and organized.
The GetKnownByAI framework: JOB × MOMENT × CONSTRAINT × FIT
A strong hint does not need to mechanically contain all four fields. The framework is a planning device to make sure your ad group describes a real decision context instead of a category noun.
Weak vs useful context hints
| Weak | Stronger | Why |
|---|---|---|
| CRM software | Sales teams replacing spreadsheets with a CRM before adding more reps | Adds job + moment |
| Running shoes | Beginner runners looking for cushioned everyday shoes for first-5K training | Adds use case + fit |
| Accounting automation | Finance teams that need invoice automation but must keep approval controls | Adds job + constraint |
| Project management | Product teams coordinating engineering and design work across weekly releases | Adds team + workflow moment |
Build ad groups around messageable intent
OpenAI recommends keeping an ad group focused on a common product category, theme or customer need. Our practical test is even simpler: could one coherent ad promise and landing page satisfy every hint in this cluster?
If the answer is no, split the group. Contexts that require meaningfully different proof, offers or landing experiences should not be forced together just because they use the same product name.
If two contexts need different headlines, different proof or a different CTA to feel relevant, they probably belong in different ad groups.
From customer research to context hints
- Collect sales calls, search terms, support questions, reviews and customer interviews.
- Extract recurring jobs, moments, constraints, comparisons and desired outcomes.
- Group contexts by shared product fit and message.
- Write natural-language hints that add information beyond the ad copy.
- Create creative variants around the same underlying intent cluster.
- Send the cluster to a landing page that proves the same promise.
What context hints should not do
- Do not stuff synonyms or long disconnected keyword lists.
- Do not describe use cases your product does not genuinely support.
- Do not mix unrelated audiences and products into one cluster.
- Do not assume an exact phrase in a hint means an ad will show whenever that phrase appears.
A reusable worksheet
| Prompt | Your answer |
|---|---|
| JOB | What outcome is the customer trying to achieve? |
| MOMENT | What changed or triggered the need? |
| CONSTRAINT | What must be true for a solution to work? |
| FIT | Why is your product a credible fit? |
| PROOF | What evidence on the landing page supports the promise? |
| CONVERSION | What action indicates this context produced value? |
Primary source
Platform mechanics verified 30 August 2026. JOB × MOMENT × CONSTRAINT × FIT is a GetKnownByAI planning framework.
Map your first context clusters.
We turn real customer language into intent architecture, context hints, ad angles and landing-page requirements.
Request a readiness sprint ↗