The news: OpenAI’s new “ChatGPT agent”, which started rolling out last week, goes beyond chatbots by acting as an autonomous “digital worker,” per TechCrunch. Available to OpenAI’s Pro, Team, and Enterprise subscribers for $200/month per user, the agent operates software, browses websites, fills out forms, and creates documents within a secure sandbox, potentially rivaling tools like Microsoft Office. Our take: As AI companies combine their models into autonomous tools, marketers, researchers, and pilots should test agents on repeatable, low-risk tasks like generating decks or summarizing reports. Exercise human oversight, track time saved, and evaluate ROI and output quality against legacy tools like Microsoft Office to determine if agents are a viable replacement.
AI is rapidly becoming foundational to marketing strategy, with 63% of teams now using it for planning—up from 28% in 2023, per Boathouse. Customer service and analytics have seen similarly sharp increases, supported by rising investments in CRM systems, CDPs, and automation tools, according to Twilio. As AI’s footprint grows, marketers are reallocating spend toward digital formats like social, CTV, and video, where AI can optimize targeting and performance. This trend reflects a broader shift: the most successful marketers are embedding AI into the fabric of their decision-making, not treating it as a plug-in. The gap is widening fast.
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: As opportunities for AI-powered ads grow, consumers remain hesitant and can even be turned off by a brand using the technology. Just 12% of US adults would be more likely to buy a product from a brand if they knew it used AI in its advertising, per CivicScience. Less than a quarter (22%) positively view brands that use AI-powered advertising, compared with 37% who view them negatively. Our take: Transparency and careful application of AI are key to avoid alienating users and build trust with consumers. Brands should introduce AI slowly by starting with prototyping ideas and generating backgrounds before diving into full-scale AI ad creation.
Over half of worldwide consumers (51%) want AI to improve their experience by helping them find products faster, according to February data from Twilio.
Last November, our analysts made some predictions about how the retail category would fare in 2025. Now that we’re halfway through the year, it’s time to check back in on what has (or hasn’t) happened. "We're seeing many of our predicted trends playing out, though not always in the ways we anticipated," said our analyst Suzy Davidkhanian on a recent episode of the "Behind the Numbers" podcast. "The retail landscape is evolving rapidly, with some developments accelerating faster than expected while others face unexpected headwinds."
AI is poised to transform everything in marketing from ad creation to targeting and even the future role of agencies in the advertising ecosystem.
The trend: Physicians are ramping up use of AI for pharma-related queries on medications, treatments, and drug interactions. But usage of AI trails search, according to a new Bain & Company. report. The big takeaway: Doctors trust search engines over AI for drug information—for now. The convergence of the two tools via AI Overviews on Google could lead to declining confidence in search results. The winner in securing physician trust could be clinical-specific AI tools like the widely used UptoDate or the emerging OpenEvidence, which brands itself as a ChatGPT for doctors.
The report: OpenAI is reportedly developing a checkout feature that would allow users to complete purchases directly within ChatGPT, according to The Financial Times. Merchants would pay OpenAI a commission on any resulting sales. Our take: At the risk of sounding hyperbolic, checkout integration could fundamentally transform the ecommerce landscape. Even before news of the feature began circulating, brands were exploring AI optimization, or “AIO”, to rank in AI-generated product recommendations. Now, with purchases just a click away, ChatGPT could emerge as a viable commerce engine—especially if it undercuts incumbent marketplaces’ take rates. And it likely won’t be alone for long. Rivals like Perplexity and Anthropic are almost certain to build similar transactional layers, creating a new crop of marketplace-like platforms for sellers in short order.
On today's podcast episode, we discuss the state of some of our 2025 predictions, including GenAI’s influence on business growth, the influence of China’s e-commerce disruptors, the squeeze on retail media networks, and more. Then, we offer a few more slightly spicier predictions for the remainder of the year ahead. Listen to the conversation with our Senior Analyst Sara Lebow as she hosts Vice President Suzy Davidkhanian and Senior Analyst Carina Perkins.
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: The window to monitor AI’s reasoning in chatbots and agents is quickly closing, according to 40 researchers from Google DeepMind, Anthropic, OpenAI, Meta, and more. In a rare show of unity, the researchers stated that chatbots and agents are shifting from human-readable chain-of-thought reasoning to opaque, non-verbal methods, per VentureBeat. Our take: The collective call for transparency and standards marks an inflection point. Without urgent action, AI systems may soon outpace our ability to audit them—leaving marketers, creators, and regulators flying blind. Unseen logic means unchecked bias that could result in reputational damage.
The news: The breakneck speed of AI development makes bugs easier to miss and slower to patch, leaving platforms vulnerable to flaws and potentially leaked data. This week, Meta revealed it had patched a bug in January that would have let its AI chatbot users access others’ private prompts and responses, per TechCrunch. Key takeaway: As AI tools become central to marketing workflows, so do the risks tied to prompt exposure, IP leaks, and client data breaches. Marketers must approach AI adoption with the same scrutiny they apply to any vendor handling sensitive assets.
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.
The news: Shopify will not allow agents and other bots to purchase on users’ behalf without “final human review,” the company said in an update to the code used by merchants to operate their online storefronts. Our take: While AI agents aren’t yet reliable enough to be given free reign over purchase decisions, companies have to be prepared for a future where they soon will be.
The news: Google is experimenting with AI summaries in Discover—the news feed within its iOS and Android search apps—adding yet another threat to referral traffic for web publishers. Instead of displaying a headline and link to a news story, Discover shows an AI summary with an icon featuring the logo of any cited source. Our take: If users increasingly rely on AI summaries—and if Discover becomes a zero-click search hub—publishers risk further declines in web traffic, imperiling not just ad revenues but the viability of good journalism.
Nextdoor is undergoing a major reinvention, focusing on hyperlocal value with three core features: real-time safety alerts, AI-generated neighborhood recommendations, and curated news from over 3,500 local publishers. The redesigned platform aims to capitalize on shifting work-from-home behavior, verified neighbor identities, and underused local advertising budgets. With 100 million registered users across 11 countries, Nextdoor is uniquely positioned to offer geotargeted content and build ad inventory through increased daily engagement. CEO Nirav Tolia’s bet? Depth over scale. If executed well, the new Nextdoor could become an essential tool for local businesses, publishers, and residents alike—while opening fresh monetization streams.
Generative AI is playing a growing role in video advertising, with 22% of all video ad creative enhanced by genAI in 2024—a figure set to reach 39% by 2026, per IAB. Smaller advertisers are leading the way, using genAI to scale affordable, personalized content. Major platforms like Meta, TikTok, YouTube, and Amazon are fueling adoption with built-in tools that boost ROAS and compress production timelines. While creative speed and flexibility are increasing, many marketers still face hurdles around measurement and platform-level data transparency. As capabilities improve, genAI is transforming video from a production bottleneck into a performance engine.
The news: AI agent adoption in business is happening at an accelerated rate with companies like Intuit, Capital One, and Highmark Health revealing how agents are solving problems and disrupting enterprise workflows, per Venturebeat. Our take: Enterprise AI agents have moved from labs to the front lines. For marketing leaders, that means a clear opportunity to start applying agents to accelerate creative work and squeeze inefficiencies out of existing workflows. As AI agent use becomes mainstream, ensuring an oversight on safety and reliability will become necessary requirements in protecting brand reputation.
The news: xAI, Elon Musk’s AI company, issued a public apology after Grok posted extremist, antisemitic, and politically incendiary content. The chatbot described itself as “MechaHitler” and repeated far-right rhetoric—shortly after Musk pushed to make the chatbot “less politically correct.” Our take: Despite Grok’s competitive performance, its volatility may keep it off the table for marketers running AI pilot programs—like NinjaPromo, which is piloting AI tools that combine its proprietary models with external LLMs for predictive analytics, generative content, and programmatic ads aimed at boosting ROI and automating workflows. Before trusting any platform, CMOs must ensure their tool’s transparency and determine how each model reasons—and what values or biases are embedded.