Retailers have built lucrative revenue streams from retail media networks (RMNs), leveraging on-site ad inventory and first-party transaction data. As the potential grows for consumers to shop through AI agents instead of retailer sites or apps, those data streams and ad surfaces are at risk.

The news: OpenAI CEO Sam Altman is warning of a growing AI investment bubble. “Are we in a phase where investors are overexcited about AI? My opinion is yes,” Altman said during a dinner with a group of reporters, per The Verge. Still, he emphasized that AI remains “the most important thing to happen in a very long time.” Our take: Altman’s warning about an AI bubble applies to marketers too. The temptation to chase every shiny new AI tool is real, but teams should develop an AI experimentation roadmap with clear outcomes to avoid wasting resources. Pushing vendors for case studies can help maximize budgets.

AI shopping assistants are reshaping product discovery and threatening retail media’s search-driven model. RMNs and advertisers need to adjust strategies to protect their market position and seize the opportunity to turn disruption into growth.

The news: OpenAI’s GPT-5 could be the start of ChatGPT becoming a transaction-driven super app that monetizes user intent, not attention. GPT-5’s router—which analyzes queries and decides how hard to “think” based on complexity—lets OpenAI invest more resources during high-intent moments like “compare hiking boots under $200” or “best smart TVs for co-op gaming.” Prioritizing queries with high commercial value could help OpenAI monetize users not through ads but via affiliate or take-rate revenues, per SemiAnalysis. Partnerships with Shopify and others suggest that monetization stack is already on the way. Our take: A full-service ChatGPT that’s intuitive enough to guide full shopping journeys inside a chatbot while keeping backend costs minimal could rewrite the AI platform’s business model. Brands should be working to optimize for AI-native commerce and integrate with agentic tools.

The news: Despite consumers’ rising use of AI agents for search, shopping, and discovery, brands are falling behind on generative engine optimization (GEO) strategies. 47% of brands have no deliberate GEO strategy or have no idea if they appear at all in AI agent responses, per a new report from Cordial. Another 47% have only just begun optimizing content for AI discovery. Our take: To boost visibility, brands should optimize for conversational context and create structured, machine-readable content that AI can index, like clear website FAQs, TL;DR summaries, and detailed product specs. Expanding presence across social platforms that feed AI training models, such as Reddit, Quora, and YouTube, can also improve chances of surfacing in AI-generated responses.

The news: GenAI models can easily be influenced to perpetuate false health facts when they’re fed made-up medical terms and information, per a new Mount Sinai research published in Nature last week. Our take: As more consumers rely on open GPT models for health answers, misinformation and intentional disinformation pose growing risks to both personal and public health. There’s an opportunity for healthcare and pharma marketers to step up science-based AI marketing and communications, such as Pfizer’s custom genAI medical query tool Health Answers that sources answers from medical journals and peer-reviewed research.

The news: OpenAI is bringing its newest models to Amazon Web Services (AWS) for the first time, marking a major milestone in the ongoing battle for AI cloud dominance. Its open-weight models, gpt-oss-120b and gpt-oss-20b—available via Amazon Bedrock and SageMaker AI platforms—are able to handle complicated text-based operations and integrate into cloud-based systems. Our take: OpenAI’s models are getting easier to access, meaning lower costs and fewer technical hurdles to trying powerful AI tools. AWS customers should start testing oss-120b and oss-20b for things like generating subject lines, social copy, and campaign variations and explore ways to fine-tune the models with company data.

The news: Amazon plans to put ads in its AI-powered Alexa+ voice assistant to boost product discovery and profits. Our take: If Amazon rolls out sponsored answers to Alexa+ user queries or in-conversation ads, the voice assistant’s vast trove of personal user data will help marketers target consumers on a micro level. However, if hallucinations arise and lead to irrelevant or inaccurate product recommendations, Amazon risks eroding both user trust and brand confidence.

The news: OpenAI abruptly discontinued a ChatGPT “share” feature after widespread criticism of its opt-in functionality surfaced thousands of unintended private chats in Google Search results. If a user checked a box to “make this chat discoverable” (sometimes accidentally or not fully understanding the warning), Google and other search engines could “see” these chat links and add them to public search results. Our take: The AI industry’s “move fast and break things” ethos is clashing with the non-negotiable demands of data protection. For marketers reliant on AI for strategic planning and analysis, security and data privacy are paramount. Companies demonstrating a strong security focus could stand out from competitors.

US digital ad spending growth will dip below 10% YoY in 2025, but retail media, social networks, and CTV will still perform above the overall average.

The news: More than half (51%) of customer service journeys start on search engines and third-party platforms like Google, YouTube, Reddit, and ChatGPT—rather than company websites—prompting businesses to meet customers where they are, per a recent Gartner survey. Our take: Brands need to research and identify the platforms their customers rely on and establish fast, responsive service on those channels. The goal isn’t to pull users back to official websites—it’s to meet them where they already are, with the answers they need, when they need them. Using generative engine optimization (GEO) best practices to boost customer service answers in genAI outputs could help younger consumers get digestible, fast answers in their preferred channel.

The news: OpenAI is preparing to launch GPT-5, a model that will combine traditional GPT capabilities with o3-series reasoning—marking a major leap in performance and model simplification. Our take: GPT-5 could streamline content creation, search, and CX workflows, leading to renewed industry adoption and customization. Enterprise customers should test GPT-5’s API early. Align adoption with marketing workflows and consider consolidating tools into a single platform to reduce costs. Early movers will shape the future of customer engagement.

The trend: GenAI tools like ChatGPT are providing fewer disclaimers that chatbots are not a substitute for professional medical advice, according to a recent study cited in MIT Technology Review. Our take: Tech players must prioritize user safety—not winning the AI race. Health warnings should be standard, and marketers will need to scale back claims that AI accuracy surpasses physicians’.

Alphabet posted strong Q2 results, with Search ad revenue up 12% YoY and YouTube ad revenue climbing 13%. But analysts and advertisers are asking tougher questions as the company shifts toward AI-led formats like AI Overviews and Gemini. Google declined to provide clear data on ROI, clickthroughs, or user engagement, fueling concerns about monetization in a no-click world. Licensing costs for LLM training, brand safety, and competition from ChatGPT and Perplexity are all in focus. While YouTube continues to lead in streaming ad growth, the future of Google’s ad engine may hinge on transparency, AI accountability, and performance parity.

Walmart is going all in on AI as it prepares for a future in which more people rely on the technology to work and shop. The company is making strategic AI hires while streamlining its agents to make them easier for shoppers, employees, and partners to use. Agentic tools are both a threat and an opportunity for retailers. Companies need to prepare for a future where tools like ChatGPT and Perplexity make purchases on behalf of shoppers—which will require them either to make their websites more accessible to these assistants, or to build their own AI agents to make those transactions seamless. AI agents are also a useful investment in the current era of uncertainty, given their ability to unlock cost savings at a time when every dollar counts.

The news: ChatGPT isn’t just leading the chatbot race—it’s dominating it. With rapid growth and billions of daily prompts, it remains the go-to generative AI (genAI) tool for both businesses and consumers despite rising competition. The stats: Figures on the AI leader’s user growth are debated, but the most consistent recent number for weekly active users (WAUs) is 500 million—an increase from 100 million WAUs in November 2023, a year after its debut. That number tripled to 300 million by December 2024. Our take: As genAI evolves from a novelty product to a routine tool in workflows and daily life, ChatGPT is maintaining its role as an industry leader. To stay ahead, OpenAI should focus on improving ChatGPT as a core product and prove it can scale profits sustainably before racing to outbuild its rivals.

GenAI will reach about 51% of US internet users by 2029 as growth stabilizes, with search dominating use cases and Gen Z leading adoption. Amid rising competition from Google and others, ChatGPT will maintain dominance. Brands must adapt to AI-mediated customer relationships.

The news: Perplexity is in talks with smartphone manufacturers to make its new Comet browser a default app on smartphones to drive adoption and user engagement, per Reuters. Perplexity CEO Aravind Srinivas said it aims to reach “tens to hundreds of millions” of users in 2026 after a desktop rollout to a “few hundred thousand” testers, a plan that could be aided by expanding Comet access on phones. Our take: While Comet itself is a browser, its integrations with Perplexity’s AI could streamline access to mobile AI search tools, changing mobile search behavior and forcing marketers to rethink traditional search marketing practices. Getting Comet onto phones could also supercharge Perplexity’s data on user behavior and boost its ability to improve its AI search tools.

The news: The vast majority of referral traffic from AI sources comes from desktop users while mobile traffic lingers in single-digit percentages. 94% of ChatGPT referral traffic is from desktop users, per BrightEdge’s The Open Frontier of Mobile AI Search report. Google Gemini’s traffic is 94% desktop versus 5% mobile, while Perplexity’s is 96% desktop and just 3% mobile. Our take: As search engines increasingly reduce organic visibility and prioritize zero-click searches, brands and publishers need to develop unique content strategies for different devices. Providing a mix of long-form, in-depth posts for desktop users along with snappy headlines and skimmable content for those on mobile could help achieve the best of both worlds

The news: OpenAI is preparing to launch a suite of office productivity tools that could let users bypass tools from Microsoft. Users will be able to build and modify presentations and spreadsheets that are compatible with PowerPoint and Excel, per The Information—without using Microsoft’s own apps. Our take: This suite could position OpenAI as a serious contender in the office software space, bypassing years of Microsoft and Google development. Companies using ChatGPT could improve workflows by managing documents, generating content, and executing repetitive tasks from a single interface.