Context hints are the plain-language descriptions an advertiser writes in ChatGPT Ads to tell OpenAI's matching system which conversations, topics, and buyer situations an ad should be eligible to appear inside. They act as relevance signals that describe a buying moment, and they neither bid on keywords nor guarantee placement in any specific chat.
If you advertise on ChatGPT, context hints are the main lever you control over who sees your ad, since there are no keywords, demographics, or in-market audiences to buy. Write them too broadly and your budget spreads across irrelevant chats; write them with precision and your ad reaches people the moment they weigh a purchase.
A context hint is a short, natural-language phrase, entered at the ad-group level in the ChatGPT Ads dashboard, that describes the user moment where your product or service is relevant. OpenAI's matching system reads the meaning of each live conversation and compares it to your hints, then decides whether your ad is eligible to appear.
A workable hint carries three things: who the user is, what they are trying to do, and the context that scopes the moment. You write it as plain English in a single freeform field, with no keyword lists, match types, or negative-keyword sheets to manage. OpenAI treats the hint as a description of intent, so the system aligns it with the embedding of an ongoing chat instead of matching literal words.
This puts context hints in the targeting slot a keyword would occupy in Google Ads, yet they behave like audience descriptions built from intent. They set relevance only; campaign settings still handle geography, platforms, and audience lists. AirOps builds and updates on-brand ChatGPT ads, including their context hints, from your Brand Kit so hint, creative, and policy rules stay consistent as you scale.
Context hints feed OpenAI's relevance system, which decides in real time whether your ad fits an active conversation. The flow runs from what you write to what the system serves.
Describe the moment: At the ad-group level, you write a hint that names the product, the buyer, and the situation in one natural phrase.
Keep groups tight: You split meaningfully different products, audiences, or use cases into separate ad groups so each hint stays focused.
Match against chats: OpenAI compares the meaning of each eligible conversation to your hints and your creative, landing page, and targeting selections.
Serve or skip: When the conversation aligns with your intent, your ad becomes eligible; OpenAI still chooses which eligible ad to show.
Read the results: Reporting shows impressions, clicks, and conversions by ad group, which tells you which hints earned delivery.
That output tells you which hints won eligibility and drove action. It does not tell you which conversations you appeared in, because OpenAI never shares chats, so you tune hints from aggregate performance.
Resources: See how AirOps approaches writing ChatGPT ads context hints, specs, and measurement
Context hints decide whether your ChatGPT Ads budget reaches buyers or burns on conversations that will never convert. Because they are the one relevance lever you control, the quality of your hints sets the ceiling on everything downstream, from click-through rate to cost per conversion.
They control your reach: Hints are the primary way you tell OpenAI which conversations qualify, so vague or off-target phrasing is the fastest way to waste real budget on irrelevant chats.
They demand a new skill: Teams moving from Google or Meta have no keywords, demographics, or in-market audiences to fall back on, so writing intent in plain language becomes a core campaign skill.
They shape measurable outcomes: Because tightly themed hints map to tightly themed ad groups, they directly affect how readable your reporting is and how fast you can find the variants that actually earn delivery.
SEO managers use context hints to extend the intent they already map for organic answers into paid ChatGPT placements on the same buying questions.
Content strategists use context hints to match each ad group to one buyer scenario, so creative and landing pages stay aligned to a single moment.
Growth marketers use context hints to test keyword-style phrasing against full-sentence descriptions and keep whichever ad group delivers cheaper conversions.
Each context hint lives inside a single ad group, so the real unit you target is the group itself, and OpenAI is explicit that meaningfully different products, audiences, or use cases belong in separate ad groups instead of one combined bucket.
Context hints shape only which conversations your ad qualifies for; they cannot control geography, schedules, budgets, or exclusions, so you set every hard limit in campaign settings and treat the hint purely as a description of relevant intent.
OpenAI matches the meaning of your hint to the meaning of each live conversation, so a hint that describes who the buyer is and what they are trying to do pulls far more qualified placements than a bare list of product words.
Reach buyers inside ChatGPT during the conversation where they describe a need.
Target by intent without keywords, demographics, or in-market audiences to buy.
Align each ad group to one buyer moment for cleaner creative and reporting.
Adapt phrasing quickly, since hints are plain English in a single editable field.
Extend your organic AEO intent work into paid placements on the same questions.
Write each hint as one natural phrase covering a product, a buyer, and a situation, because a described moment matches conversations better than a word list.
Keep ad groups tightly themed and split different products or audiences into their own groups, so each hint and its creative stay focused.
Point every ad group at a landing page that matches its hint, since mismatched pages waste the qualified clicks the hint earns.
Set geography, platforms, and audiences in campaign settings, because hints cannot enforce any hard limit on where ads run.
Test keyword-style phrasing against full-sentence hints in separate ad groups, and keep whichever delivers cheaper qualified conversions.
Allow OAI-AdsBot to crawl your landing pages, because a page the crawler cannot load stalls your launch.
Avoid stuffing one ad group with hints for several unrelated products or audiences. OpenAI reads blended intent poorly, so the matcher spreads your ad across loosely related chats and your reporting stops telling you which hint actually worked.
AirOps: builds and updates on-brand ChatGPT ads and their context hints from your Brand Kit, and shows where your ads and competitors' ads appear across the prompts you track.
OpenAI ChatGPT Ads dashboard: the self-serve interface at ads.openai.com where you write context hints at the ad-group level and set campaign targeting.
Google Analytics 4: measures what visitors from your ChatGPT ad landing pages do after the click, so you can judge which hints drive real conversions.
Map the moment: Write down the exact buyer, product, and situation for your first campaign in one plain sentence. You can do this today with no budget approval.
Draft tight hints: Turn that sentence into one context hint per ad group, keeping each group to a single product-and-buyer combination. Describe the situation in plain language so the matcher can read intent, and avoid stuffing several products into one group.
Prepare the landing pages: Allow OAI-AdsBot to crawl each page, and match every ad group to a page that fits its hint. A page the crawler cannot load will stall your launch, so check robots.txt and firewall rules first.
Set the hard limits: In campaign settings, choose your countries, platforms, and any audiences, since hints cannot control these. Add the United States explicitly if you want US-only delivery.
Launch and compare: Run three to five ad variants per group, then read impressions, clicks, and conversions to find which hints earn delivery. Change one element per variant so each result has a clear cause.
Context hints are plain-language descriptions of the buyer moments where an advertiser wants a specific ChatGPT ad to appear.
You write them at the ad-group level as natural phrases that name a product, a buyer, and a buying situation.
Hints influence relevance only and cannot control geography, schedules, or exclusions, which belong in campaign settings.
Blending unrelated products or audiences into one hint spreads spend across loosely related chats and muddies your reporting.
The biggest gains come from tightly themed, intent-rich hints paired with matched landing pages and steady, disciplined testing.
Context hints describe a buyer and a moment, while a Google Ads keyword tries to match the literal words someone types. A keyword lives in a bidding system with match types, negative keywords, and exact strings the engine aligns to a query. A context hint has none of that machinery: you enter one natural phrase in a single field, and OpenAI compares its meaning to the meaning of a live ChatGPT conversation. Because the match is interpretive, the same hint can qualify for many differently worded chats that share the same intent, and it can also miss chats that use your words but mean something else. The practical shift is that you stop thinking about strings and start thinking about people, describing who the user is and what they are trying to accomplish so the matcher can recognize the moment when it appears.
Use as few context hints as it takes to describe one clear buyer-and-product combination, and move anything meaningfully different into its own ad group. OpenAI does not publish a strict limit, and the useful rule is thematic focus instead of volume: one ad group should represent a single intent so your reporting stays readable and the matcher is not pulled across unrelated conversations. In practice that means you add hints only when they describe the same moment from slightly different angles, and you create new groups the instant a hint would cover a different product, audience, or use case. When you launch, start with three to five ad variants in that focused group so you can learn which creative works against the hint. Review delivery after the group has spent enough to show a pattern, then tighten or split hints based on which conversations earned your ad.
Delivery shifts because context hints set eligibility instead of a fixed placement, and the conversations happening on ChatGPT change constantly. Your hint describes an intent; whether it gets matched depends on how many people are having conversations that fit that intent on a given day, how your creative and landing page score, and which other eligible advertisers are competing for the same moment. OpenAI chooses which eligible ad to serve, so even a strong hint will see natural swings in impressions as demand and competition move. Personalization adds another source of variation, because when a user turns it on, signals from their broader ChatGPT experience can influence the match. The honest read is that day-to-day movement is expected, so judge a hint over a window long enough to smooth out the noise, and do not react to a single slow day.
No, you cannot directly control which conversations your context hints appear in, and it helps to plan around that from the start. OpenAI keeps final control over which eligible ad is shown and in which chat it appears, and advertisers never receive the conversations themselves. What you can influence is eligibility and relevance: the precision of your hint, the match between your creative and landing page, and the campaign settings that gate where ads can run at all. So the right move is to make your intent unmistakable and let the system do the matching. Write a hint that captures one clear moment, point it at a page that delivers on that moment, and use countries, platforms, and audiences to rule out placements you never want. Then read aggregate performance and refine, since that feedback loop is the real lever you hold over placement.
A good context hint reads as one natural sentence that names a specific product, a specific buyer, and the situation that brings them together. The clearest test is whether a stranger could read your hint and picture the exact person and moment you are trying to reach; if the phrase could describe several unrelated audiences, it is too broad. Strong hints avoid keyword stuffing and stay scoped to a single intent, because that focus is what lets the matcher connect your ad to the right conversations and keeps your reporting legible. There is no universal score to hit, so judge quality by outcomes: a hint is working when it earns steady impressions on relevant chats and drives conversions at an acceptable cost. If delivery looks broad or costs climb, tighten the wording or split the group, and compare a sentence-style hint against a tighter variant to see which one your buyers actually respond to.