AI Agents Have IDs Now

It all started during a client call last week. We weren’t talking about AI. We were discussing fraud detection, identity checks, and some compliance headaches piling up across their customer service flows. Midway through the conversation, someone asked, half-joking, “When AI starts answering support tickets on our behalf, who’s going to vouch for it?”

It wasn’t meant to be a deep question. But it stuck.

Because for all the excitement around autonomous agents and copilots doing real work writing emails, handling inquiries, analyzing documents we’re glossing over something fundamental: how we manage their identity.

Telecoms, have spent decades building infrastructure around verifying and managing human identities. SIMs, eSIMs, device authentication, KYC protocols, we know this terrain. But we’re entering a new zone, one that doesn’t behave like the old one. These AI agents don’t walk into a store, show a passport, or pass a biometric scan. They live in systems. They move fast, spin up dynamically, operate across services. And yet… they’ll be transacting. Accessing data. Speaking on your company’s behalf. Making decisions.

The old model wasn’t designed for that.

What’s become clear, especially in conversations with telcos and tech leaders over the past few months, is that we’re about to hit a wall. Everyone’s excited about deploying AI agents into workflows. Fewer are thinking about how we’ll manage trust when hundreds of those agents are running across different functions, vendors, or even countries. Who owns their identity? Who revokes it when it goes rogue? And how do you prove one agent is authentic while another is a hallucinated clone?

It’s not a future problem. It’s already creeping into current systems. Some teams I’ve worked with are piloting internal copilots that pull from multiple CRMs, analyze customer history, and generate responses. They’re fast, impressive, and in many cases, completely untracked in terms of who did what. The audit trail is fuzzy. The accountability is murky. And the identity management? Usually a shared login or internal token.

That’s not going to hold.

This is where I think telcos have a bigger role to play than they realize. Most people still think of eSIM as a tool for traveler convenience or device provisioning. But at its core, it’s a programmable identity layer, secure, scalable, and already embedded into millions of devices. The same infrastructure we use to manage phones and IoT endpoints could become the trust fabric for AI agents. Not in a sci-fi way. In a practical, infrastructure-meets-security kind of way.

And yes, it would require rethinking how we issue credentials. How we define “identity” for an agent that isn’t tied to a physical user. But the shift is already underway. If agents are going to act with delegated authority, sign contracts, retrieve data, trigger workflows, we’re going to need to authenticate them with the same rigor we apply to human users.

It reminds me of early IoT conversations years ago. Everyone was racing to connect devices, fridges, sensors, lightbulbs, but no one had a solid plan for securing them. The mess that followed wasn’t a surprise. We’re now standing at a similar crossroads. Agents are being connected before they’re being protected.

What’s missing isn’t the technology. It’s the conversation.

We need a serious rethink of digital identity and this time, not just for people. Enterprises will need identity management systems that can handle thousands of AI-powered entities, each with different roles, access levels, and operational risks. Telcos already understand how to do this at scale. The infrastructure is there. The mindset shift isn’t. Yet.

This isn’t about locking things down. It’s about building the kind of trust that lets us move faster. Because when you know who or what, you’re dealing with, you stop wasting time on verification. You can actually scale.

I’m still working through the implications. I don’t think anyone has all the answers yet. But the question that keeps circling is this: if AI agents are going to be part of our teams, systems, and customer interactions, who is going to issue them their badge?

And if telcos don’t step into that role… who will?

---

If you want to know more about how Apiro Data can help your Digital Transformation, IoT and operational strategy.

Get in touch, and let’s build something great together!

Or email us at info@apirodata.com

Apiro Data is Your Guide to Navigating the Digital Landscape. Our versatile modular solutions guide your business smoothly through digital transformation and operational enhancements. Our experience-driven approach delivers tangible solutions grounded in real-world wisdom.

Follow us on LinkedIn.