Do Users Actually Want Ads in ChatGPT?

Amazon bringing ads into ChatGPT allows us to give a core judgment upfront: e-commerce advertising has entered conversational decision-making flows for the first time. Whoever occupies the position before the user voices their purchase intent is closer to closing the deal.

Having worked on IoT platforms for several years, I'm used to looking at processes first. When a user asks a question, the system needs to understand intent, recall options, compare parameters, give suggestions, and ideally jump straight to ordering. Previously, search and recommendation pages did this. Now ChatGPT absorbs the middle layer into the conversation.

The materials say advertisers can place ads in ChatGPT via Amazon DSP, with the pilot initially limited to the US. Campaigns are set up and managed by Amazon, while placement decisions are controlled by OpenAI's ad system. Ad units appear as text or images below ChatGPT responses. This division of labor is subtle. Amazon provides merchants, budgets, and placement tools; OpenAI provides user scenarios and final display rights.

Ads will never influence the answers ChatGPT gives; answers are optimized to provide maximum help to users.

This sentence is the premise for whether this business model can stand.

From a user experience perspective, the question is direct. If a user asks what air purifier is suitable for a bedroom, and the AI gives an answer followed by sponsored products below, do users need this scenario? It might be useful. The most annoying thing in conversations is insufficient specificity—model numbers, prices, and compatibility conditions often aren't clear. If ads can bring up comparable options, saving users from searching again, that's good. But it cannot change the answer itself. Once recommendations are influenced by budget, AI's credibility slides from "helping me judge" to "helping me sell."

This is also the tension between Amazon and OpenAI. Amazon needs AI traffic but doesn't want OpenAI freely scraping its product data. Product libraries, prices, reviews, and inventory are its core assets. The ad partnership looks like opening a controlled channel: "I can give you budget and placement interfaces, but I won't fully open my shelf data."

I've been tinkering with gateways and smart home devices at home recently. After less than two months, I have deep feelings. Device ecosystems can't just look at devices being connected. Interface permissions, status feedback, and installation barriers determine whether users continue using them. ChatGPT ads face the same issues. Integrating with Amazon's ad system is just the first step; whether users can smoothly view products, compare parameters, and place orders after clicking is what completes the transaction flow. If the process breaks between display and purchase, forcing users back to search pages to order, the value becomes thin.

For advertisers, they are buying the moment of decision. Traditional display ads buy exposure, search ads buy intent, and conversational ads buy the segment where intent is most complete. Users may have already stated their budget, usage, and home environment. If ads can connect here, conversion rates will be higher than generic traffic. According to public reports, ChatGPT ad pilots have expanded to more markets, with some platforms mentioning related revenues reaching billion-dollar levels. The direction is clear: OpenAI needs commercialization, and Amazon needs new traffic.

However, the greater the commercial value, the harder it is to maintain product purity. Independent answers, independent ads, and clear labels are trust maintenance mechanisms. Short-term, users can accept this. Long-term, platforms will face pressure to make ads look more like recommendations, slowly shifting independent displays to natural appearances. Smart homes have seen this process too: starting with just status displays, later automation, push notifications, memberships, and accessories all crammed into one entry point.

So, the Amazon and OpenAI partnership is probing ownership of an entry point: does shopping decision-making happen in conversations or in e-commerce Apps? Ads are just probes. If users get used to completing comparisons in ChatGPT, Amazon must continue providing transaction interfaces; if users only treat AI as advice and transactions still return to Amazon, then OpenAI's ad value remains merely traffic diversion.

Whether users continue to trust AI depends on whether platforms can separate answers from ads.