The End of Apps? Inside Google's Bet on an Agent-First Future

This week we sat down with Lucrezia Noli, Head of AI for Healthcare & Life Sciences at Google — and her view of where this is all heading is bolder than most. The apps you tap through today, she argues, are scaffolding for something that's about to be torn down.

🎧 Short on time? Hit play and let the conversation run while you get on with your day — listen to the full episode here.

The pitch is simple to say and hard to absorb: we've spent fifteen years learning to navigate apps — tapping, swiping, hunting through menus. Lucrezia's case is that agentic AI flips the relationship. Instead of you operating software, you state an intent and a layer of agents goes off and operates the software for you, stitching together whatever tools and data the job needs. The interface stops being a screen full of icons and starts being a conversation.

What makes that more than a slide-deck future is grounding — wiring large language models to real-time, trustworthy data so their answers reflect what's actually true right now rather than what was true at training time. Sat on top of Google's Gemini and the early work on world models, that's the difference between a clever chatbot and an agent you'd let act on your behalf.

The part that should land hardest for anyone leading a team is what Lucrezia calls the displacement paradox. Agents are getting genuinely capable, yet human judgement isn't going away, it's moving up the stack. The work shifts from doing the task to deciding which tasks are worth doing, checking the agent's reasoning, and owning the call when it matters. Multi-agent set-ups, where specialised agents collaborate and hand off to each other, only sharpen that: someone still has to design the system and trust it under pressure.

Why it matters: if the agent-first shift plays out the way Google is betting, the winners won't be whoever builds the slickest app — they'll be whoever gets their data clean, grounded and trustworthy enough for an agent to act on. That's a strategy question, not an IT one.

👉 Listen to the full conversation with Lucrezia Noli — including her take on AI marketplaces, where world models beat LLMs, and the surprising link between jazz and AI.

You don't send a colleague a three-word brief. You explain the context, the constraints, what you've already tried. But typing all that into ChatGPT takes forever — so you don't.

Wispr Flow lets you speak your prompts instead. Talk through your thinking naturally and get clean, paste-ready text. No filler words. No cleanup. Just detailed prompts that actually get you useful answers on the first try.

Millions of users worldwide. Works system-wide on Mac, Windows, and iPhone.

Try Wispr Flow free

Enjoyed this? Project Flux lands in your inbox every Monday with the week's most important AI and tech news, decoded for people who actually have to make decisions. 👉 Subscribe to the Project Flux Newsletter — it's free.

Project Flux is independently owned, funded, and operated by its founder. Any external references or features are for community benefit only and do not imply endorsement, control, or ownership.