In partnership with

We track Product so you don't have to. Top Podcasts summarised, the latest AI tools, plus research and news in a 5 min digest.

Hey Product Fans!

Welcome to this week’s 🌮 Product Tapas!

If you’ve been forwarded this or just stumbled upon it, you’re in for a treat. For the best reading experience, check out the web version and sign up for future editions here.

What’s on the menu this week? 🍤 

OpenAI launches an entire platform stack in one week (once again—everything everywhere all at once), agents graduate from chatbots to browser-controlling digital employees (your cursor has a cursor now), and A16Z data reveals copilots beat autonomous agents because humans still want to feel useful. Meanwhile, a tiny Caribbean island funds half its budget selling .ai domains (best passive income strategy yet).

• 📰 Not Boring → OpenAI's platform blitz, agent takeover accelerates, enterprise AI scales up

• ⌚️ Productivity Tapas → QA automation for voice agents, knowledge-to-AI-agent tools, monetisable workflow builders

• 🍔 Blog Bites → Trust-revenue attribution engines, AI agent orchestration tactics, Patagonia's Earth ownership model

• 🎙️ Pod Shots → How Surge hit $1BN+ revenue without raising a dollar (quality over everything)

Let's go 🚀

📰 Not boring

Rise of the Agents

• Cursor can now control your browser; take screenshots, improve UI, and debug client issues and Google's Gemini 2.5 Computer Use (snappy name) model can now navigate websites and is supposedly better than 'leading alternatives' (compelling…)

• Google's Jules brings coding agent into terminals, Slack and other tools as AI coding agent competition heats up

• ElevenLabs introduces Agent Workflows. Visually build conversation flows and route to subagents to handle complex cases

• Nothing launches AI tool for building mini apps using prompts

We are very much in the "agents everywhere" phase it seems. When your browser becomes controllable by AI and Google's pushing computer-use models that can navigate websites, we're not talking about chatbots anymore—we're on the road to digital employees. The race to build the most capable agent is heating up fast, with everyone from coding tools to hardware companies jumping in. The question isn't whether agents will be everywhere, but how quickly they'll make current workflows obsolete.

OpenAI's Platform Play Gets Serious

• OpenAI has launched apps inside ChatGPT, so you can draw in content from a variety of sources

• They also launched AgentKit to help developers build and ship AI agents (taking a big swipe at automation tools like Zapier and N8N)

• Plus ChatKit - providing a simple embeddable interface that developers can use to bring chat into their own apps and Evals for agents tools to measure AI performance

• Codex (their AI coding tool) is available to all ChatGPT Plus, Pro, Team and Enterprise users

OpenAI isn't just building better models—they're building the infrastructure to become the operating system for AI applications. AgentKit directly targets Zapier and N8N's automation market whilst ChatKit makes it trivial to embed AI into any app. This is classic platform strategy: make it so easy to build on top of OpenAI that developers have no reason to go elsewhere. The distribution advantages are massive too when you already have hundreds of millions of users.

The Data Grab Intensifies

• Meta will start using your chatbot interactions (1bn MAU) as training data

• As part of OpenAI's e-commerce announcement, Stripe announced an open standard for how agents can interact with payments (Agentic Commerce Protocol-ACP). You can read more about it here

• Perplexity has made its Comet browser free for everyone

Meta's billion monthly active users chatting with AI bots represents quite the training dataset. Meanwhile, Stripe's creating the payment rails for the agent economy, and Perplexity's giving away browsers to capture more search behaviour. Everyone's positioning for the next phase where data quality matters more than data quantity—and user interactions with AI systems are the highest quality data you can get.

Enterprise AI 

• Deloitte will make Claude available to 470,000 people across its global network and IBM has integrated it with a stack of its applications

• A16Z released their AI Application Spending Report: Copilots lead the way (ahead of autonomous agents), OpenAI is at the top followed by Anthropic. Vibe-coding and creative tools are unsurprising fast followers. Amazed Granola isn’t there though.

• Replit announced Connectors; a way to build apps and automations that seamlessly integrate with the tools you already rely on every day

When Deloitte's rolling out Claude to nearly half a million employees, we're past the pilot phase. The A16Z data is interesting: copilots are winning over autonomous agents because they augment rather than replace human decision-making (although how long will that last?). Replit's connectors approach is another obvious trend—instead of building everything from scratch, make it trivial to connect AI to existing workflows.

Odds and Ends

• You can now use Nano Banana inside Google Slides to edit images and Videos

• Anthropic is bringing Claude Code to mobile

• Facebook updates its algorithm to give users more control over which videos they see

• Shein is opening its first physical stores, in France

• Aaaaaand finally, the tiny Caribbean island of Anguilla now derives almost half of its state budget from the sale of .ai domain names

Google's putting AI image editing directly into Slides whilst Anthropic's making Claude Code accessible on mobile—the integration into everyday tools continues. Shein's physical stores are interesting—even digital-native brands need real-world presence. And Anguilla's .ai domain windfall is the most unexpected AI beneficiary story of the year.

Learn AI in 5 minutes a day

This is the easiest way for a busy person wanting to learn AI in as little time as possible:

1. Sign up for The Rundown AI newsletter

2. They send you 5-minute email updates on the latest AI news and how to use it

3. You learn how to become 2x more productive by leveraging AI

Sign up to start learning.

⌚️ Productivity Tapas: Time-Saving Tools & GPTs

• Built for Mars Figma plugin: Login to your Built for Mars account, and use the "Insights" tool directly from your Figma workspace.Select a frame and get instant UX feedback and ideas. View relevant sources directly from the BFM+ library. Click to drop post-it notes directly into your board

• Glazed: creates and ships tracking events from Figma files

• Integrity: Bring notes canvases and AI chats into one connected workspace. Miro + Claude+Notes

• Pencil: agent driven MCP canvas built around open design format that lives in your codebase. Basically natural language design for IDEs like Cursor [waitlist]

  Remember. Product Tapas subscribers get our complete toolkit - 460+ personally tailored, time-saving tools for PMs and founders. Your shortcut to efficiency and what's hot in product management 🔥

Check the link here to access.

🍔 Blog Bites - Essential Reads for Product Teams

Growth: The Trust-Revenue Engine That Transforms Content into Cash

Whilst marketing focused, there’s plenty to take away from this recent Got Catalyst article. In it, Will explores the critical attribution gap plaguing fintech marketing teams who generate impressive engagement metrics but struggle to prove actual revenue impact. He provides a systematic framework for connecting content consumption directly to closed deals through proper tracking architecture. Read the full piece here

💡 "Most fintech companies can't prove which content drives revenue, when it drives revenue, and how much revenue it drives."

Product Development: From Code Monkey to AI Orchestra Conductor

This one isn’t for everyone… but of you want to go a bit deeper into AI agents and Agent swarms read on…! In a recent blog post, Zach Wills explores a revolutionary development approach that transforms engineers from linear coders into orchestrators of parallel AI agent swarms. He demonstrates how managing 20 simultaneous AI agents can compress weeks of development into days, fundamentally shifting the role from implementation to intelligent direction. Read the full article here.

💡 "The AI wasn't the bottleneck. I was. My workflow, built around a single thread of attention, was the thing holding back progress."

Strategy: When Earth Becomes Your Only Shareholder

Tom from Strategy Breakdowns explores Patagonia's revolutionary $3 billion ownership transfer that rewrote the rules of purpose-led business. Rather than maximising shareholder value, founder Yvon Chouinard created a radical structure that legally requires all profits to fund environmental causes in perpetuity. Read the full article here.

💡 "Instead of extracting value from nature and transforming it into wealth, we are using the wealth Patagonia creates to protect the source… We're making Earth our only shareholder. I am dead serious about saving this planet."

🎙️ Pod Shots - Bitesized Podcast Summaries

Remember, we've built an ever-growing library of our top podcast summaries. Whether you need a quick refresher, want to preview an episode, or need to get up to speed fast - we've got you covered.

Check it out here

🎯 How Surge Hit $1BN+ Revenue Without Raising a Dollar

A per last week I sent out the full Pod Shot summary earlier in the week so as to keep this main newsletter shorter.

🎥Watch the full episode here

📆 Published: July 25th, 2025

🕒 Estimated Reading Time: 8mins. Time saved: 60 mins!🔥

Link below to the full summary - but here’s what’s in store:

"Quality is the most important thing. It's more important than anything else." — Edwin Chen, CEO of Surge

Edwin built a $1BN+ revenue company with no funding, no sales team, and 10% of the headcount of funded competitors. How? By ruthlessly cutting the 90% of work that doesn't matter.

Key insights from the full interview:

• ⚖️ 90% of Big Tech work is useless — small, focused teams move exponentially faster

• 🛑 No standing 1:1s — if you need scheduled meetings, your real-time communication is broken

• ⚡ Build MVPs in weeks, not months — modern tooling makes this possible for 95% of products

• 🎯 Quality is non-negotiable — it's the only sustainable moat in AI data labelling

• 🧠 100x engineers exist — AI amplifies exceptional talent far more than average

• 🔬 Data quality > compute — frontier labs waste 6-12 months on contaminated data

• 📊 Popular benchmarks are broken — LM Arena rewards formatting over accuracy

• 🔮 Multiple AGIs will coexist — different models for different strengths, like mathematicians vs. poets

👉 Read the full breakdown — sent separately, check your inbox

That’s a wrap.

As always, the journey doesn't end here!

Please share and let us know what you would like to see more or less of so we can continue to improve your Product Tapas. 🚀👋

Alastair 🍽️.