A week in AI is like a year in other industries. I hope these issues become your weekly source of AI information, inspiration, and ideas. If we haven’t met before, I’m Amanda Smith. I write about AI and the fascinating folks who are building in this brave new world.
AI agents are starting to shop, book, and pay on our behalf, and the systems businesses rely on to keep transactions safe weren't built for them.
This week's founder is building the trust layer for this new kind of customer.
Give it a read if you're building an AI-native product.
This week in AI:
Tech leaders sign AI agreement
President Trump and leaders from leading AI labs to sign a voluntary safety framework.
Anthropic releases Sonnett 5.5
30% faster and costs 30% less.
Muse privacy complaints (already)
Without user consent, of course.
🤝 This edition is kindly brought to you by Vistage
You don’t have to go it alone. With a confidential circle of trusted CEO peers, an accomplished coach and a proven approach, you can lead with clarity, confidence and better decisions.
We partner with a select group of brands that we use or admire, allowing this newsletter to remain free and independent. Reach out for your campaign here.

Company background: Beltic
Founded: January 2025
Team size: 8 + a collection of agents and advisors
Funding to date: $8.8 million ($1.5M pre-seed led by Collide Capital, with Latitud and Positive Ventures, and a $7.3M seed led by Norwest, with Restive Ventures and Oxford Seed Fund)
ARR: Not disclosed
Isha Bhatnagar: Beltic
Isha Bhatnagar spent six years at leading cryptocurrency exchange, Coinbase, as the head of product for regulatory experience – building the identity and verification infrastructure that 50+ million users interact with. While she held roles at Netflix, Twitter and Accenture, it was Coinbase that most influenced her entrepreneurial journey, as she was spearheading nascent technology.
What became clear to Bhatnagar was that what got in the way of identity scaling was the infrastructure of the system, not the new technology.
“Because the financial system is deeply fragmented, signals of safety are weak, and the design is optimized for operational volume and requires a human in the loop at every step,” she explained.
“If that doesn't scale for humans today, it definitely can't scale for agents. When AI agents started to come on the scene, I recognized that the same infrastructure problem I'd spent years on was going to be a real barrier to a huge economic opportunity. The customer just wasn't human anymore.”
Bhatnagar met her Beltic co-founder through a mutual friend, and they went to Internet Identity Workshop, “an unconference of sorts that is one of my favorite Silicon Valley gatherings.” Every conversation kept coming back to agents and it was clear identity was about to shift in a big way.
For decades, risk management came down to asking "are you human?' That question was not only about to stop being useful, but it was also going to be a blocker.
Why should agents have an identity?
“Agents are transacting now, not in two years – for both businesses and consumers,” Bhatnagar said.
“So, while every fraud system was built for humans (looking at selfies, device fingerprints, mouse movement, click speed, etc.), the agents hoping to transact don't leave those traces. Companies either block them and lose the customer or accept them blind and take on the risk.” Neither works, she said.
Beltic was built to answer the questions a business needs when an agent shows up at checkout: Who's behind it? Is it allowed to do what it is trying to do? Is this what the person actually wanted?
“Businesses need to treat agents as first-class traffic, governed by similar policies they apply to people. If you don't let a customer book more than two tables a night, their agent shouldn't either. The companies that figure this out first will win the customers who've stopped clicking through websites.”
The build & go-to-market
Bhatnagar explained an agent can behave perfectly on one action and completely differently on the next, so a one-time check doesn't work. Beltic assesses every action against the user's intent and the business's existing rules.
“One key decision was owning our whole technical stack. We acquired Verifiet in 2025, which brought global entity data infrastructure (and it’s found) in house. U.S.-focused providers tend to stop at U.S. data, and agents and the people behind them don't stop at the border. Owning the data, the context, and the engine is what makes our decisions instant and defensible to a regulator. You can't get there stitching vendors together.”
Beltic started out servicing the companies that need verification the most: Financial institutions, card issuers, payment providers, and large merchants. Long term, their customer is any business receiving agent traffic.
“Say your agent books your flight, hotel, travel insurance, and a ride to the airport. Every one of those companies needs to serve you confidently, under its own rules.”
While Beltic isn’t sharing revenue yet, they did reveal they’re working with a select group of design partners at the forefront of agentic commerce and payments.
Scaling plans & predictions
Given its early days, growth is founder-led, and they’re present in every customer conversation. Bhatnagar said the most underrated channel has been small, in-person rooms, as well as the caliber of advisors and investors who have connected them to big companies.
Beltic’s design partner program also works well as a growth strategy. “Companies get early access and help shape the platform, so they're invested from day one,” she said.
In the next year, Bhatnagar predicts we’ll see businesses move from blocking agents to serving them, with guardrails they control. Identity standards for agents will start to consolidate, and the question will become whether you can trust what an agent is doing when it tries to transact with your business.
We will move from user-initiated actions to agent-initiated actions.
What will this look like?
“Agents will call other agents and hand work to systems they don't own. Every handoff resets the trust assumptions, and inherited trust without verification is just assumed trust. Beyond payments, we'll see agents taking high-risk business actions on their own, and that's where our solution matters most,” she concluded.
How's the depth of today's edition?
If one of these stories stuck with you, I’d love to hear which one.
Speak soon,
Amanda
P.S. If you want to get a founder feature about your own story, reply to this email. If you’d like to reach our newsletter audience (founders, creators, and marketers), click the button below.
If you’re new here, I’m over the moon you’ve joined us! To help me craft content that’s actually useful (and not just noise in your inbox), I’d love it if you took 1 minute to answer this quick survey below. Your insights help shape everything I write.
✨ Insane Media is more than one voice
Dive into our other newsletters - where psychology meets the creator economy, e-commerce marketing, and Human resources.









