Ads are moving into chatbot conversations, where they can draw on more context than a conventional search query. That may help advertisers match a message to what someone is discussing, but it also raises questions about privacy, relevance and trust. The early market is growing, though its shape is still uncertain.
Chatbots are becoming another ad channel
Anthropic says it will keep Claude ad-free, a pledge it made in its Super Bowl commercial this year. Its recent IPO filing did not mention selling ads as a way to fund the chatbot. The article notes that Anthropic has about $112 billion in investor cash, giving it room to hold that position.
Other companies are testing advertising as a source of revenue to help cover the high costs of generating AI responses. OpenAI has attracted over 100,000 customers advertising in ChatGPT since ads debuted there in February. Google has expanded ads in its search chatbot AI Mode to 10 more countries beyond the US, and says users “find them highly relevant.” Advertisers, however, are not yet satisfied with the experience, even as they expect it to improve.
Smaller chatbot platforms may have a stronger need for ad income because they cannot count on the same scale of venture capital as top AI labs or the revenues of tech giants. Koah Labs says it brings advertising to over 200 apps, including DeepAI and Liner, and works with brands such as Hims and UrbanStems.
More context can change how an ad fits
Koah CEO Nic Baird says ads composed partly by AI generate double the clicks of ads written entirely by people. The article offers a possible explanation, rather than a settled cause: AI-written marketing copy may feel more natural beside AI-generated responses.
Koah bases ads on product details and images supplied by advertisers, which Baird says helps avoid inaccurate information made up by AI. Still, some advertisers, including drugmakers, must follow government requirements about what their ads can say. Baird says those companies tend to keep more human control over the copy, and that their ads perform less strongly.
Chatbots may also support a different kind of targeting. A search for “help with stress” might lead to an ad for a therapist, while Baird says spa ads have performed well after prompts about “stress.” He argues that Koah can consider the prompt, the AI response and the broader conversation when choosing an ad.
Chatbot prompts tend to be at least double the length of Google queries, according to the article. Koah says users click on about 3 percent of its ads, roughly in line with the more mature Google ads system. Those figures suggest potential, but they do not settle whether users welcome ads in a conversation.
Personalization brings a privacy trade-off
When Koah selects an ad, it generally receives the user’s full conversation up to that point. That can reveal more than a single search phrase, especially because people may use chatbots as confidants as well as information tools. Whether users accept that level of context being used for personalized ads remains unclear.
Baird says Koah anonymizes queries and stores them in a de-identified way, but declined to give further details about its data collection and storage. He argues that relevant context could also shape the page a user sees after clicking an ad. For example, the page might address a product feature the person mentioned or fill in a downside that the chatbot did not discuss.
Koah does not have data showing whether people want these customized pages. Conversation history could also help advertisers show the same user an ad again, a practice known as retargeting. The potential benefits depend on users seeing the personalization as useful rather than intrusive.
Support moments and shopping agents remain open questions
Koah estimates that one out of every four chatbot queries about a retailer or brand involves a return, exchange or other support issue. Baird sees a possible opening for companies to use ads to offer warranty information or AI-assisted support to disappointed customers. Ideally, that would help someone resolve a problem faster; a poorly placed ad could instead waste money and make the customer more frustrated.
The apps Koah currently serves do not offer AI agents that shop independently. But Instinct and Meta’s Muse are starting to take over shopping chores for some users. Meta CEO Mark Zuckerberg has said Muse could become the company’s biggest source of revenue, suggesting transaction fees rather than ads as the source.
Ads could still influence AI agents, the article says, though any sponsored material would need to be clearly identified to protect user trust. The market could also stall if more developers follow Anthropic’s ad-free approach. DeepAI CEO Kevin Baragona says ad revenue has been meaningful, but he is not committed to it because the experience does not yet feel right on privacy, design and other grounds. The central challenge is not simply placing an ad in a conversation; it is making the placement useful enough to justify the context it uses.