Three channels. Senior on every call.
People use ChatGPT to compare options and decide what to do next. OpenAI now sells ad space below those answers. We set up your account, wire the tracking first, and run a measured test, so you learn whether the channel pays before you scale it.
ChatGPT ads show below an answer, matched to what the person is asking about. OpenAI says people are often comparing options in those moments, and there are no keywords to buy. The work is in the signal: tracking, context, and creative.
Your business owns the OpenAI Ads Manager account and the billing. We get a role inside it, help you through business verification, and install measurement before a single ad runs.
There are no keywords to bid on. You describe when your ad is relevant through context hints, and OpenAI matches it to conversations. We write those hints and enough ad variations for the system to find the right moments.
OpenAI, GA4 and your CRM will not show the same numbers, and OpenAI says so too. We put them side by side, label modeled conversions as modeled, and judge the channel on leads you can trace.
Four phases. Qualify, wire, test, decide. A new channel earns a bigger budget by proving itself against the channels you already pay for.
Is your category allowed? OpenAI currently limits ads mainly to consumer categories and approves some others case by case. We check your category against its ad policies, plus your landing pages and current tracking, before anything is built.
Account, verification, Pixel, lead events, UTMs and GA4. We test every event fires before any ad goes to OpenAI for review.
A controlled budget on a Clicks campaign, with several ad variations and context hints per service. Once enough leads come through, we test a Conversions campaign.
We compare ChatGPT cost per lead to your search and social numbers and make a plain call: scale it, hold it, or stop spending.
ChatGPT Ads is new, and we do not have a ChatGPT Ads case study yet. We will not invent one. These three files come from the same tracking-first method on Google Ads. Open HVAC for the full replay, or see every case study.
Ads spending more, earning less. We surfaced broken conversion actions, fixed the tracking, restructured the campaigns. Same business. Vastly better numbers.
Open the replayDashboard said "winning," chairs said "empty." Made the tracking honest. Conversion rate doubled. Bookings number reads like a typo. It isn't.
See all case studiesManufactured home developer needed broader reach without burning budget. Lower CPC, lower cost, more conversions, more pipeline calls.
See all case studiesChatGPT Ads is OpenAI's ad platform, bought through Ads Manager at ads.openai.com. Everything below comes from OpenAI's own documentation. It is still a beta, so we re-check the details before every launch.
The ad sits below a ChatGPT response, clearly labeled and kept separate from the answer. Each unit carries your brand name, your logo, a title, a short description, an image and a link to your landing page. OpenAI states that ads do not influence what ChatGPT answers.
Ads show to people on the Free and Go plans. Plus, Pro, Business and Enterprise users do not see them, and neither do accounts identified as under 18.
There are no keywords to buy. OpenAI's system weighs the conversation and what the person is trying to do, your landing page, your title and copy, and the context hints you write for each ad group. When a user has ads personalization turned on, it can also use select signals from their wider ChatGPT use. A relevance-weighted, second-price auction then picks among the eligible ads.
Advertisers never see the chats. Reporting comes back aggregated: impressions, clicks, spend and conversions, not prompts. Want to appear inside the answer instead of below it? That is AI search optimization, and no ad budget buys it.
Every campaign runs on one objective, and the objective cannot be changed after the campaign is created. OpenAI bills per 1,000 impressions or per valid click. You never pay per conversion.
Billed per 1,000 impressions. Built for visibility, not leads, so it says little about what a customer costs. For a business that runs on leads, we start with Clicks instead.
Billed per valid click. OpenAI suggests starting with a maximum bid of $3 to $5 per click, or letting its Maximize results strategy set bids. This is where our tests start: real visitors, real landing-page behavior, a first cost per lead.
The system optimizes toward one supported standard event from OpenAI conversion tracking, billed per click or per 1,000 impressions. One event per campaign, and it cannot be swapped later, so we only move here once tracking is proven.
Budgets behave differently too. A daily budget is an average across a Sunday to Saturday week, so a single day can spend up to twice the daily figure. The floor is $25 a day in US dollars. Card-billed accounts are charged each time unpaid spend reaches a payment threshold, plus any balance left at month end. We set budgets with that swing in mind.
The path OpenAI documents for a new advertiser, in order. Half the work happens before a single ad exists. We handle each step with you.
Sign in at ads.openai.com with an OpenAI login and create the advertiser account. Country, currency and time zone cannot be changed once it exists, so they get set right the first time.
Ads Manager keeps three names: account, brand and legal. The brand name is what people see beside the ad. Billing and verification must be complete before anything delivers.
OpenAI's OAI-AdsBot crawler has to reach your landing pages for ad review. Firewalls, bot protection and robots.txt rules can block it. We test access before any ad is submitted.
OpenAI Pixel through Google Tag Manager, a supported standard event for leads or bookings, UTMs on every ad URL, and a test that each event actually fires.
Objective, budget, dates, countries and platforms. To run in the US only, the US has to be included on purpose. Excluding other countries does not limit delivery.
One ad group per service or customer need, each with its own context hints and either a max bid or OpenAI's Maximize results bidding.
Several titles, descriptions and images per ad group, each a different angle, each pointing to the most relevant page rather than the homepage.
Every ad passes OpenAI review first. Once live, clicks report in about 15 minutes, spend can lag 7 to 8 hours, and conversions can take 24 to 48 hours.
On ChatGPT you describe the customer and the moment, and the system looks for matching conversations. Specific beats clever, every time.
OpenAI's guidance is to say what you offer, who it helps and when it is useful, one idea per hint, in plain language. "Water heater replacement for older homes that have lost hot water" tells the system far more than "water heater."
Hints shape relevance only. They cannot enforce a service area, a schedule or an exclusion, so those limits go in the campaign's own delivery controls, where OpenAI offers them.
OpenAI favors volume and variety: several distinct titles and descriptions per offer, each from a different angle, written to be useful rather than catchy. The landing page should match the ad instead of defaulting to the homepage, and images should stay simple and on message.
Every ad is reviewed against OpenAI's policies, and misleading or inconsistent ads get rejected. We write to that standard from the first draft.
Different auction, different moment. ChatGPT Ads adds to search and social. It does not replace them. We run all three inside one paid ads lane, judged on the same cost per lead.
Your ad appears below an answer while someone works through a question. Matching runs on the conversation and your context hints. OpenAI states there are no performance benchmarks for the platform yet.
You bid on the words people type, with match types and negative keywords you control. Years of data and deep reporting. For most service businesses it stays the core of paid search.
Ads reach people in their feeds based on audiences and interests, not on what they asked. Strong for creating demand and for retargeting, with creative doing most of the work.
An honest filter. Being early only pays if you can measure what you bought.
No bait-and-switch to a junior after sign-on. The senior who took the first call is the senior who runs the work, with the rest of the bench behind them when the lane needs it.
▮ The team
Every lane is owned by a senior who's run the work for a decade. Every account gets the same discipline: the conversion taxonomy is honest before any dollar moves, and the senior who explained it on the first call is the one watching it daily. The bench supports the senior, not the other way around.
Tell us what you sell and where your leads come from today. We will tell you honestly whether ChatGPT Ads is worth a test for your business, and what a fair test budget looks like.
Book the call
Tell us what you sell and where your leads come from today. We'll talk it through with you and come back with an honest read on whether ChatGPT Ads fits, plus a test plan if it does.