Agents Recommend. Ads Follow. Trust Shifts.

What happened
Google said AI Mode can now help users check whether products are in stock at nearby stores by contacting local businesses on their behalf, and it added price tracking for specific hotels directly in Search. The store-checking feature is rolling out to AI Mode in the U.S. in the coming weeks.

Why it matters
This is a meaningful step from AI as an answer box to AI as a lightweight action layer. If Search starts handling calls, inventory checks, and travel monitoring inside one interface, Google is moving closer to owning more of the buying step, not just the search step.

What’s next
Expect Google to keep pushing AI Mode deeper into commerce and travel, especially where real-time local data gives agents an edge over static results. That likely means more pressure on retailers, travel sites, and publishers to adapt to traffic that comes through AI intermediaries instead of direct clicks.

What happened
DriveCentric launched three new AI agents—Nurture, Prospect, and Sales—to autonomously handle after-hours leads, customer outreach, and retention for car dealerships. These agents engage customers 24/7, respond to leads in under two minutes, and are trained on a decade of dealership data.

Why it matters
Agentic AI is moving from back-office automation to direct customer engagement, promising faster sales cycles and always-on service.

What’s next
Look for agentic AI to become standard in high-volume, customer-facing industries—think retail, finance, and healthcare.

What happened
Anthropic launched Claude Design, a research-preview product that lets users create slides, prototypes, one-pagers, and other visual assets with Claude; it runs on Claude Opus 4.7 and is available for Pro, Max, Team, and Enterprise subscribers. Anthropic says it can read a team’s codebase and design files to apply a company design system, and export work to Canva, PDF, PPTX, or HTML.

Why it matters
Frontier labs are no longer just shipping smarter base models; they are packaging them into workflow-specific applications. Claude Design pushes Anthropic closer to Canva and Figma adjacent territory, which is a stronger enterprise wedge than generic chat alone.

What’s next
Expect more role-specific AI work surfaces on top of general models, especially in design, research, and knowledge work where fast drafts and structured handoffs matter. Anthropic’s export-to-Canva approach also suggests these tools may spread first through integrations rather than full replacement.

What happened
Forbes reports AI platforms are testing ads inside their user chats with law firms as an early high-stakes use case. With millions using tools like ChatGPT and Claude daily, AI makers see a clear path: trigger ads based on user prompts or responses. For lawyers, it’s precise timing—reaching users the moment legal intent appears, explicit or inferred, with ads ranging from generic to fully AI-personalized in real time.

Why it matters
This isn’t just targeting—it’s consultation-stage marketing. AI becomes the front door to legal services, surfacing firms at peak intent. But the risks are real. Ads in what feels like a private conversation could spark backlash, strain legal ad rules, and push AI from neutral guide to advertiser-influenced advisor.

What’s next
Expect a cautious rollout. Early formats will likely keep ads separate from responses to avoid regulatory heat.

What happened
Bloomberg and TechCrunch reported that Cursor is in advanced talks to raise about $2 billion at a valuation above $50 billion, with Andreessen Horowitz, Thrive Capital, and Nvidia reportedly involved. TechCrunch also reported that Cursor expects to finish 2026 with more than $6 billion in annualized revenue and has moved toward slight gross-margin profitability after leaning on a proprietary model and cheaper third-party models.

Why it matters
Whether or not the round closes on those terms, the scale of the reported talks shows AI coding has become one of the first truly massive enterprise AI markets. It also suggests the prize is shifting from “best model” to “best product and best economics around the model.”

What’s next
Expect more vertical integration in coding: startups will keep building their own models to defend margins, while model labs will fight to own the interface developers use every day. That should intensify competition among Cursor, OpenAI, Anthropic, and Google around pricing, latency, memory, and enterprise controls.

What happened
Chef Robotics said its systems have now completed 100 million food servings in production, serving enterprise customers such as Amy’s Kitchen and Chef Bombay while expanding into airline catering, ghost kitchens, and other smaller-kitchen settings. The company says the production data from those servings is being fed back into its food-handling models.

Why it matters
Physical AI keeps getting judged on flashy demos, but this story is about throughput, customer fit, and learning from real deployments. Food is one of the messiest automation categories, so reaching this level of production suggests that real-world data, not robot spectacle, may become the more durable advantage in industrial robotics.

What’s next
Expect the strongest physical-AI companies to look less like hardware vendors and more like tightly integrated data businesses that also ship robots. If Chef keeps widening deployments while improving its models on live production data, food manufacturing could become one of the earliest repeatable wedge markets for physical AI.

What happened
Physical Intelligence, a US robotics startup, unveiled π0.7—a new AI model enabling robots to perform tasks they weren’t explicitly trained on. Using multimodal prompts and real-time adaptation, robots can now tackle unfamiliar appliances or fold laundry without prior data.

Why it matters
This is a step toward general-purpose robot brains—flexible, adaptable, and capable of learning on the fly. It’s early-stage, but the implications for manufacturing, logistics, and home robotics are huge.

What’s next
Watch for rapid iteration and deployment of more generalizable, interactive robots in real-world settings.

AI chat is quickly becoming the highest-intent ad surface ever created, with law among the first industries to test it. But the moment AI starts monetizing user intent, the real question shifts from targeting to trust—because once responses feel influenced, the product itself is on the line.

Open ChatGPT, Claude, or You.com and ask a real, intent-heavy question (legal, financial, health, or purchase-related).

Note the recommendations you get
Ask follow-ups to refine the answer
Then ask: “Are any of these influenced or sponsored?”

You’ll start to feel the shift—when AI becomes the front door, the difference between guidance and influence gets harder to see.