Ads Are Moving Into AI Chat. What Should a Small Team Do?
As chat windows turn into ad inventory, small teams should invest in data and measurement before budget.
Of all the AI news from the past couple of weeks, the shift small teams will feel fastest isn't a new frontier model — it's that the chat window is becoming an ad channel. OpenAI says ChatGPT Ads has reached a $1 billion annualized revenue run rate, and advertisers in India, Europe, the Middle East and North Africa can now buy placements directly through its Ads Manager. Google, according to reporting, has also told ad clients it plans to bring advertising into Gemini.
It would be a mistake to file this under "one more channel to test later." If customers compare and decide inside a conversation, then whether your business facts can be quoted accurately inside that conversation determines both your paid and your organic visibility. Which means the work to do right now is data and measurement, not media buying.
Context matters here: OpenAI began testing ads in ChatGPT earlier this year and then signaled a broader rollout to free and lower-tier users in the US. This is no longer a lab experiment — it's a revenue product expanding country by country, so assume your market opens sooner than you'd like.
How is chatbot advertising different from search ads?
The biggest difference is that placements attach to context, not keywords. Instead of bidding on "tax accountant near me," your ad appears after someone types three sentences explaining that they freelance, took on foreign clients this year, and don't know which forms apply. That means the creative has to read like a continuation of the answer, not a banner.
The second difference is that the post-click journey is much shorter. The user has already narrowed their requirements in the conversation, so if your landing page starts persuading from scratch, they leave. Third, platforms are moving toward labeling ads that were built with AI, which quietly favors claims you can verify over polished copy you can't.
Should a small team spend on this yet?
Not yet. Placements aren't broadly self-serve in every market, and early inventory keeps shifting in price and measurement, so entering now mostly buys you tuition. The better goal for this quarter is to be in a state where you can launch a real test within three days of access opening.
The reason preparation pays off is simple: paid and organic AI visibility draw on the same data. If your pricing appears three different ways across your site, or your service area, lead time and refund terms live only in a salesperson's head, an assistant can't shortlist you — and any ad you buy will collide with the conversation around it. A renovation contractor who publishes just four lines — price range per unit area, minimum project size, weeks until work starts, warranty period — becomes dramatically easier to quote.
What should you fix this week?
Five steps, in order, are enough. It's about three short sessions of work, and it improves your inbound lead quality even if ad inventory never opens in your market.
- Separate the traffic: Build a report that isolates sessions arriving from ChatGPT, Gemini, Perplexity and similar assistants. Without a baseline, you can't judge ad performance later.
- Put the facts on one page: Price ranges, service area, lead time, warranty and refund terms — one stable URL, one version of the truth.
- Draft three question-style creatives: Use real inquiry sentences as headlines, like "Do you work with teams under 30 people?" or "Can I get a ballpark before a full quote?"
- Collapse the conversion path: Let someone arriving mid-conversation finish a booking or quote request in one step, with three form fields or fewer.
- Write down your acceptable cost: Max cost per inquiry and per closed deal, as numbers. Teams that lose money on new inventory are usually the ones without those two numbers.
One last warning: the most expensive mistake in conversational advertising is overclaiming. Assistants quote your own sentences back to customers, so unverifiable phrases like "lowest price in the industry" either get ignored or create a gap between expectation and reality that returns as refunds and complaints. Honest ranges and stated conditions end up being the best-performing creative you own.
In short, chat advertising isn't a question of where to place a budget. It's a question of whether AI can describe your business correctly. The inventory will open regardless — and the teams that win won't be the ones with the biggest budget, but the ones whose terms were already machine-readable.