Human in the Loop

This morning I keep coming back to something simple. For all the technology we build, the systems that last are the ones that keep a human in the loop.

The last couple of weeks I ran into email deliverability issues with Monday Thought. Some of you never saw the welcome email. Others missed a couple of editions entirely.

Not in the inbox. Not in spam. Just blocked. It stung, because I put care into every message and want them to land with you.

After some back-and-forth, the fix turned out to be a backend detail: re-authenticating DKIM records, did the trick. Technical, yes. But underneath it, the reminder that trust depends on traceability.

Email works because you can tie a message back to a sender. Finance works because you can tie transactions back to an identity.

And AI will only work at scale if we can tie outputs back to real people who are accountable.

The future is not about removing humans.

It is about deciding where to place them, and how to design for their judgment, their accountability, their presence. Without that, trust dissolves.

With it, we move forward.👇

⚡ Quick Hits

• 🧠 AI Agent Payments Protocol (AP2) launches — Google and industry partners just introduced a protocol for AI agents to execute payments on behalf of users using cryptographically signed mandates. Why it matters: it’s a real test of how we can safely let agents act while keeping humans anchored. (This is rolling out now) Investors.com: Google shakes up digital payments with AP2

• 🗂️ Projects in ChatGPT (shareable workspaces) — OpenAI now lets you create and share “Projects” with teammates. Why it matters: it bakes collaboration and accountability into AI use, keeping human context front and center. OpenAI Help Center

• 🔐 India’s new digital payments rules emphasize risk-based authentication. The Reserve Bank of India recently finalized guidelines allowing additional checks based on transaction risk, beyond baseline 2FA. Why it matters: we’re seeing policy move toward more dynamic authentication. Reuters: India central bank allows risk-based checks

• 📇 Federal AI project registry in the works . The federal government is planning a public registry of its AI projects to bring more transparency to how AI is used in government services. Why it matters: it’s a move toward making human oversight and accountability visible. Global News: Public registry for federal AI projects

🛠 Tool

🎨 Google Mixboard — A new experiment from Google Labs that lets you turn prompts into visual moodboards. You can start with a single phrase or idea and Mixboard pulls together a canvas of images, styles, and concepts to explore.

Why I like it: it makes multimodality tangible. Instead of staring at a blank screen, you can generate a board, adjust it, and then guide it with your own judgment. It’s a reminder that tools are most powerful when they expand your choices but still keep you in the loop.

Try Google Mixboard 👀

🪨 The Pebble

This week, notice where trust in your work comes from. The systems might run on code, but the real weight comes from the moments you stay in the loop. Look at the week ahead and ask:

👂 Where does your presence make the difference between something just working and something worth trusting?

— Alejandro Jaegerman

📊 Poll

✉️ If someone forwarded this to you, you can subscribe at MondayThought.com.

[#21] 👋 Feeling a little anxious we are basically already in October which is calendar Q4. Making bets and building momentum towards 2026. Maybe the year of the human in the loop? Thanks for reading → Have an amazing week!

🫆Quick ask: I’d love to learn more about you. It only takes 2 minutes, and it helps me shape Monday Thought around what matters most to readers like you. Take the survey.