// use case · B2C lead gen
Before asking for a quote, a consumer tells ChatGPT about their house, budget and timeline. Your form turns up inside that conversation, just as the project takes shape, and every lead is followed through to the appointment held.
my 1990s house runs on oil heating, should I start with a heat pump or with insulation?
The order depends on the building envelope. A heat pump in a poorly insulated house will be oversized and expensive to run. Have the loft and walls assessed first, then size the heating. An energy audit gives you the order of works and the grants you can claim.
Home energy audit
Free quote, an advisor calls you back.
your-installer.com
// why this use case
Only signed-in adult users on the Free and Go plans see ads, which is a fair description of a consumer preparing a project. In B2B, part of the target pays for a subscription and stays out of reach; here, that caveat weighs far less.
Before they know which professional to call, a consumer lays out their situation ("1990s house, oil heating") and asks where to begin. A context hint catches exactly that moment, well before a keyword comes into play.
Leaving a phone number with a stranger has a cost, and the ad and the landing page are there to lower it by saying who calls back, when, and what the person gets in return. Nothing more.
A completed form says nothing about the value of the lead. It becomes qualified once the person picks up, confirms the project and books an appointment, and those three steps are what we measure, not just the first.
// intents
Typical prompt
"I want to insulate my house before winter, where do I start and what grants exist?"
What the ad has to do
Lay out the order of steps, then offer a visit or an audit: the ad carries the first step, never the whole job.
Typical prompt
"What is a 65 m² flat with a balcony in Nantes worth, if I sell within six months?"
What the ad has to do
Offer the valuation of someone who knows the neighbourhood. The form is limited to the address and the timeline.
Typical prompt
"Online accounting course for a career change, which one do employers recognise?"
What the ad has to do
Carry verifiable proof (certification, outcomes) and lead to a conversation with an advisor rather than a generic brochure.
Typical prompt
"My father needs a hearing test, how does it work and where do we go?"
What the ad has to do
Reassure on the process and open online booking. This category falls under OpenAI's advertising policies, so eligibility is checked before any launch.
// context hints
Here a hint tells the story of a homeowner, their house and the decision they are hesitating over, and an ad group gathers the hints of one life situation: the works, the grants, the choice of installer. The excerpt below shows how a home energy retrofit installer's account could be structured:
Generate your context hintsIllustrative excerpt. The real structure is built from your search terms, sales calls and support tickets.
AG · Renovation project
AG · Grants and financing
AG · Choosing an installer
// scenarios
Three examples that tie a conversation to an ad, a page and a metric. They are not delivery promises.
// measurement
OpenAI's native report goes no further than the click. A B2C lead, though, is judged on the callback. So we wire the OpenAI Pixel and then the Conversions API to lead_created and appointment_scheduled, connect the CRM or booking tool, and it is that whole chain, read ad group by ad group, that gets compared with the accounts in the Hiints Index.
Cost per qualified lead
Not cost per form. A lead is qualified once it has been reached and the project confirmed.
Contact rate
The share of leads reached by phone. When it drops, the fault usually sits in the form or the callback delay, rarely in the ad.
Callback delay
Time elapsed between the form and the first call. It is the most underestimated variable in cost per appointment.
Conversion to appointment
Lead reached, appointment booked, appointment held: three distinct rates, read ad group by ad group.
Cost per appointment held
The number you set against Google Ads and Meta over the same period. Everything else is intermediate.
CPC and CTR vs lead gen accounts
OpenAI publishes no benchmark, so your CPC is hard to read alone. The Hiints Index sets it against comparable accounts, which the native report cannot do.
// honest qualification
Yes, clearly
You sell consumers a service that goes through a quote, a valuation or an appointment, and the value of a signed customer justifies a human callback. If someone calls leads back the same day and you accept a short-form landing page, you have the right profile.
Yes, with one caveat
Your category touches health, regulated finance, insurance or legal: it falls under OpenAI's advertising policies and eligibility is checked before any launch. Same caution if your leads are handled by a network of agencies without a shared CRM, in which case we will measure the form, not the appointment.
Not yet
Your leads wait several days before anyone calls back, or your monthly media falls short of €3,000. Either way, the channel will produce forms rather than customers. Better to tell you now than to leave you on your current channels.
Free setup, done in under 48h. Then 15% of ad spend under €5,000 of monthly media, 10% above. See pricing →
// questions
Yes. Reach is not the limiting factor: ads go to signed-in adult users on the Free and Go plans, which is the profile of a consumer preparing a purchase. What decides the result is how fast someone calls back. Only sensitive categories, which fall under OpenAI's advertising policies, need a check beforehand.
Short. The form exists to get a callback; qualification happens afterwards on the phone and is read in the CRM. Every extra field costs leads without making the rest any better.
lead_created when the form is sent and appointment_scheduled when a slot is booked, first through the OpenAI Pixel and then through the Conversions API. After that comes the lead status in the CRM (reached, qualified, appointment held, signed), which is what really counts.
Campaigns are set per country, and no finer. Context hints describe situations, not areas: "only show in Lyon" is not a hint. The area is stated in the ad and on the landing page, and out-of-area leads are filtered in the CRM.
On cost per appointment held and signing rate, over the same period. CPC and cost per form mislead: a cheaper lead that never picks up ends up costing more.
Enough to read a contact rate, and the exact number depends on what a signed customer is worth: the more it is worth, the more you can afford to pay for a lead. Setup is free and the fee follows your media, 15% up to €5,000 a month and 10% beyond. Below €3,000 of monthly media, we advise against testing.
// keep reading
How ChatGPT Ads Works: Placement, Auction and Reporting
Where ads appear in ChatGPT, who sees them, how they are selected, and what the native report does and does not show. Updated October 2026.
Read →ConversionChatGPT Ads Landing Pages: The Page That Continues the Conversation
The landing page is one of OpenAI’s ad selection signals. One page per family of intent, the checklist, three prompt → page pairs and message match.
Read →TrackingMeasuring ChatGPT Ads Conversions: Pixel, API, UTM and CRM
The native report has six columns and no conversions. OpenAI Pixel, Conversions API, a UTM convention, events by business model and CRM reconciliation.
Read →// shall we talk?
Thirty minutes, live. We place your category in the Hiints Index and give you a straight answer, either the channel is worth a test or it is not.
Book a call30 minutes · no commitment · zero pitch
