Short answer
A good AI B2B ads strategy starts from the questions buyers ask AI, not from audiences. Map 30–60 questions, tier them by intent (problem, category, vendor, decision), put roughly 60% of budget on vendor and decision questions, run one ad group per question, write each ad as the next answer, send the click to a page that continues the conversation, and name the company after the click. Then use what you learn to improve your organic AI visibility on the same questions.
Start from the question, not the audience
Every other B2B channel starts with who: job titles, account lists, lookalikes. AI assistants don’t let you choose who, and they don’t need to. The buyer tells the assistant the problem, the company size, the industry, the tools they use and what they want to achieve, often in one message. Your strategy is choosing which of those conversations to appear in.
That’s the idea behind Conversational Intent Advertising: ABM tells you who to focus on; the question tells you who’s ready.
Step 1: map the questions
Collect 30–60 questions from five places:
- Sales calls: how prospects describe the problem in the first five minutes.
- Review sites: what reviewers compared you with and why.
- Support and onboarding tickets: the triggers that made customers look.
- AI answers: run your category questions in ChatGPT, Google AI and Copilot and note who gets named.
- Search data: long queries from Search Console are often AI-style questions.
Write each one as a buyer would, with context: “best SIEM for a 200-person fintech with a two-person security team”, not “SIEM software”. How to map buyer questions.
Step 2: tier the questions by intent
| Tier | Example question | Starting budget share |
|---|---|---|
| Problem | How do we shorten our month-end close? | 15% |
| Category | What close management software is there for mid-size companies? | 25% |
| Vendor | Which close software works best with NetSuite for 300 people? | 35% |
| Decision | Vendor A vs Vendor B for a multi-entity group: pricing and set-up time? | 25% |
These splits are our starting point, not a rule. After four weeks, move budget to whichever tier brings the most companies that fit your profile.
Step 3: choose the platform mix
| Platform | Role in the strategy |
|---|---|
| ChatGPT ads | The core: you choose the questions with context hints |
| Google search campaigns | Ads can appear in AI Overviews through broad match, Performance Max or AI Max; no separate control |
| Microsoft Advertising | Copilot placements, plus LinkedIn profile targeting on search |
| Warms up the buying group and retargets identified accounts | |
| Organic AI visibility | Reaches buyers on paid AI plans who never see ads |
Details by platform: AI B2B ads platforms compared.
Step 4: write the ad as the next answer
The buyer has just read a neutral answer. Your ad should say the one thing the answer couldn’t: the specific fit for their situation. ChatGPT recommends titles of 16–24 characters (50 maximum) and copy of 32–48 characters (100 maximum), so there’s room for one claim and one proof.
- Weak: “Leading close software. Book a demo.”
- Strong: “Close in 5 days on NetSuite” / “Built for multi-entity groups. See pricing.”
Ten worked examples: AI B2B ads examples.
Step 5: build continuation pages
The click carries context the buyer already gave the assistant. Don’t throw it away on the homepage. For each vendor or decision question, send the click to a page that opens with the answer to that question: the comparison, the integration, the pricing, the case for a company of that size. Landing pages for B2B ChatGPT ads.
Step 6: identify and route companies
Most B2B visitors never fill in a form. Company identification names the businesses behind visits, scores them against your ideal customer profile and sends the good ones to sales, with the question that brought them. That question is the opening line of the sales email. Turning account signals into meetings.
Step 7: close the answer loop
Ads show you which questions bring the right companies. Those are the questions where you most need to be named organically, because buyers on ChatGPT Business and Enterprise never see ads. Feed the winners into your AEO and GEO work: a page that answers the question, a comparison AI can cite, mentions in the sources AI reads.
A worked budget: $3,000 a month
| Campaign | Questions | Monthly budget |
|---|---|---|
| US · desktop · vendor and decision | 6 ad groups | $1,500 |
| US · all devices · problem and category | 4 ad groups | $900 |
| UK · desktop · vendor and decision | 3 ad groups | $600 |
At our planning range of $4.50–6.50 a click, that’s roughly 460–670 clicks a month. Expect a few weeks of learning before you judge any one question.
Questions
How many questions should an AI B2B ads strategy start with?
Between 8 and 15 ad groups, chosen from a map of 30–60 questions. Enough to compare, few enough to fund properly.
Should we bid on competitor names?
You can’t buy competitor names as keywords in ChatGPT, but you can write context hints for comparison conversations. Keep claims fair and supportable; OpenAI’s ad policies restrict misleading comparisons.
Where does LinkedIn fit in an AI B2B ads strategy?
Before and after: LinkedIn warms up the buying group, and retargets accounts identified from AI ads. CIA vs LinkedIn ABM.
How often should the strategy change?
Budget moves weekly, ads monthly, the question map quarterly.
Facts checked against the sources below on 15 October 2026. AI ad platforms change quickly: if something here no longer matches what you see, tell us and we will update it.
Sources
- Demand Gen Report: G2 survey of B2B software buyers (March 2026)
- OpenAI: Ads campaign targeting
- OpenAI Help Center: Write context hints for ChatGPT Ads
- OpenAI Help Center: Create ads for ChatGPT Ads
- OpenAI: Ad policies (v1.6, September 2026)
- PPC Land: Google expands AI Overview ads to 11 countries
- Index Lab: What ChatGPT ads really cost in 2026
