Agentic Browsers, OpenAI’s Pivot, and Google’s Latest Move

In this edition: from security questions around agentic browsers to OpenAI’s corporate transformation and fresh research from Wharton, we’re seeing AI’s next chapter take shape.

This is Q&AI, our blog series aimed at keeping you in the know on updates in the rapidly evolving world of AI. Sometimes, these will be quick updates on new developments in the field. Sometimes, they’ll be tips on tactics, features, or functionality. 

If you haven’t met me yet, hi: I’m Jen Taylor, CI’s Director of AI Strategy & Implementation, and your (very human) AI BFF. AI is moving at the speed of light, so I’m here to let you know what matters most now. 

Q: What’s the deal with agentic browsers like ChatGPT’s Atlas? 

A: They’re the next evolution of browsing—where your browser doesn’t just find information, it acts on your behalf. Think booking travel, summarizing pages, or managing tabs automatically.

But before you download one, there are real issues to watch.

Atlas, ChatGPT’s new browser, is raising eyebrows for how it handles data. The fine print suggests user browsing behavior could train models—and users themselves may be responsible for what’s learned from their sessions. That’s a big red flag for privacy.

The even bigger concern? Prompt injection attacks. Malicious websites can hide invisible instructions that trick the browser into sharing data or performing unintended actions. There are currently no robust safeguards against this.

Bottom line: Avoid using agentic browsers with personal logins or sensitive data for now. It’s an exciting direction, but we’re still in experimental territory when it comes to security and privacy.

Q: And what’s Google’s new tool, Pomelli?

A: Pomelli is Google’s latest generative AI content tool. It analyzes your brand tone and messaging, then generates posts or captions to match.

After testing, though, the results feel a little thin—strong on speed and structure, but short on creative spark. It’s built more for social output than storytelling.

Still, it’s worth watching. Pomelli signals Google’s intent to compete more directly in the branded content space, potentially layering into Workspace and Search down the line.

Q: What’s happening with OpenAI becoming a Public Benefit Corporation?

A: OpenAI’s restructuring means it can now eventually go public and raise capital through the stock market. It’s a strategic move that gives them flexibility to grow—but it also deepens the tension between their original mission (“build AI safely and share its benefits broadly”) and the commercial realities of scaling a multibillion-dollar enterprise.

This feels like a turning point from idealism to capitalism. The question now is whether OpenAI can balance public benefit with investor expectations.

Q: Did Wharton’s 2025 AI Adoption Study reveal anything interesting?

A: Wharton’s third annual enterprise AI study captures a shift “from hype to discipline.” AI isn’t a side project anymore. It’s operational.

Key insights:

• AI is mainstream: 82% of enterprise leaders use it weekly.
• ROI is measurable: 72% are tracking outcomes, with most seeing positive returns.
• Custom builds are rising: Nearly a third of AI budgets now go to internal R&D.
• Education gap is growing: Leaders say AI enhances skills, but training confidence is dropping.
• C-suite ownership is here: 60% of enterprises now have a Chief AI Officer or equivalent.

In short, AI is mainstream—but people and processes are the new bottlenecks. The next wave of progress will come from change management, not just new models.

MORE TO COME!
Jen

Have a question for a future edition? Submit it here!