The next ad a Snapchat user sees may not look like an ad at all. It will look like a chat.
Snap Inc. has begun rolling out a format called AI Sponsored Snaps that places brand-controlled AI agents directly inside the Chat tab, the same space where users exchange messages with friends. When a user taps on one, they are not watching a video or scrolling past a banner. They are talking to a chatbot built, trained, and paid for by an advertiser. The format was announced in spring 2026 with on-the-record support from Snapchat President Ajit Mohan, who described it as “the future of ads,” and Experian Vice President Steve Hartmann, who endorsed the product from the brand side.
For Snapchat’s more than 400 million daily active users, the majority of whom are between 13 and 34, the shift is significant. The chat inbox has long been the app’s most personal space. Now it doubles as ad inventory, and the chatbot is the ad.
How AI Sponsored Snaps work
According to Adweek’s reporting on the feature, AI Sponsored Snaps appear as prompts from a brand’s AI agent alongside friends’ messages. A user can tap to ask questions, browse products, or request personalized suggestions. The conversation can lead to off-platform actions like clicking through to an advertiser’s website or downloading an app.
The format builds on an existing ad layer inside Snap’s My AI chatbot. Snap’s own privacy disclosure for My AI describes earlier placements as “contextual ads” provided by advertising partners. When the system detects commercial intent, it shares the user’s query along with contextual signals, including age range, country, language, and operating system, with the relevant advertiser. AI Sponsored Snaps appear to extend that data-sharing model from a single chatbot into the broader messaging experience, giving brands a persistent conversational presence rather than a one-off ad card. Snap has not publicly confirmed whether the data pipeline is identical, but the structural parallels are clear.
What researchers are flagging
Two academic preprints published in April 2026 offer early, independent analysis of the risks that conversational ad formats create. Neither study examines Snapchat specifically, but both address the exact mechanics Snap is now commercializing.
The first, “Ads in AI Chatbots? An Analysis of How Large Language Models Navigate Conflicts of Interest” on arXiv, lays out a framework for identifying when a language model’s commercial obligations conflict with a user’s informational needs. The researchers argue that sponsored outputs inside chat interfaces can quietly distort recommendations in ways that static display ads cannot, because the user perceives the interaction as a conversation rather than a pitch.
The second, “Commercial Persuasion in AI-Mediated Conversations” on arXiv, goes further with experimental data. That study found that sponsored labeling affects user perception but that concealed persuasion within conversational agents can go largely undetected by participants. The implication is direct: if users cannot reliably tell when an AI is selling to them, standard disclosure labels may not be enough.
Both papers are preprints and have not yet undergone formal peer review, but they represent structured, independent research that no party to the ad transaction has an incentive to produce.
The regulatory gap
No U.S. regulator has issued guidance that specifically addresses conversational ad formats powered by large language models. The Federal Trade Commission’s existing rules on endorsement disclosures and native advertising were written for static content: blog posts, sponsored articles, influencer videos. They were not designed for AI agents that adapt their language in real time based on what a user types.
That does not mean regulators are entirely absent from the AI advertising space. The FTC’s 2024 Operation AI Comply enforcement sweep targeted companies making deceptive claims about AI products, and the agency has repeatedly signaled that existing consumer-protection law applies to AI-driven practices. But enforcement actions to date have focused on misleading marketing about AI, not on AI systems that are themselves the marketing. Snap’s format sits in that gap.
The arXiv preprint on commercial persuasion discusses disclosure limitations but stops short of prescribing regulatory remedies. Whether current law already covers AI chat ads or whether new rules are needed remains an open question that neither Snap nor the FTC has answered publicly.
What Snap has not disclosed
Snap has not released engagement metrics, A/B testing results, or conversion data for AI Sponsored Snaps. The company’s announcement includes performance and scale claims, but no independent audit or third-party measurement has confirmed those figures. Without that data, it is impossible to know whether users treat these AI agents as helpful tools or dismiss them as intrusions into a space they associate with personal messaging.
Direct feedback from Snapchat’s predominantly young user base is also missing. The academic research examines persuasion dynamics in controlled experimental settings, not on Snapchat itself. Brand testimonials from the announcement, such as Hartmann’s comments, reflect advertiser enthusiasm but reveal nothing about how a 17-year-old experiences an AI sales agent sitting between messages from friends.
Snap is also not the only company pushing AI into messaging. Meta has integrated AI assistants across Messenger, Instagram, and WhatsApp. Google has expanded Gemini into its Messages app. But Snap’s approach is distinct in one important way: it is not offering a general-purpose AI assistant that occasionally surfaces ads. It is placing advertiser-built agents directly into the chat feed, making the brand the conversationalist from the start.
What this means for users right now
For anyone who encounters an AI brand agent in their Snapchat inbox, the practical reality is straightforward. Any message from a sponsored AI is an ad, regardless of how conversational it feels. The data shared during that exchange extends beyond the words typed to include age, location, and device information. And the AI on the other end of the conversation was built to sell, not to advise.
Treating those interactions with the same skepticism applied to any sales pitch is the most grounded response available. Independent measurement, regulatory clarity, and user research have not caught up with the format. Until they do, the company rolling out conversational ads remains the primary narrator of their success, and the people raising structural concerns lack the platform-specific evidence to quantify harm.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
