Trusted Agentic AI for Access Management

As enterprise networks expand across locations, vendors, clouds and operational domains, managing access policies has become a genuine operational risk, not just an IT chore. Permissions and configuration rules evolve inconsistently across systems, teams and vendor domains. A policy can be correct in one system and outdated in another. Add a growing population of human and machine identities that all need access to network capabilities, and periodic, paper-based compliance reviews simply can't keep up.

As Steffen Krippner of Vodafone put it: "The issue is not simply whether a configuration can be changed. The issue is whether every access decision and every configuration action can be checked against current policy, contract and regulation before it creates risk."

The Catalyst brings AI-driven compliance directly into the network rather than treating it as a downstream audit function. On a running system, it demonstrates a full chain: a regulatory obligation is interpreted by an AI Compliance Agent into a candidate policy, a human reviews and approves the resulting rule change inside a Digital Identity Management (DIM) layer, the access decision is enforced, and a tamper-evident audit record is created automatically.

Zero Trust principles run through the whole design: policy dry-runs before anything goes live, human-in-the-loop review at the critical decision point, controlled rollout, and test data injection to validate behaviour safely. The work is grounded in TM Forum standards, specifically TMF720 and TMFC020, so it functions as a practical, standards-based blueprint rather than a one-off demo.

Karsten Thon of Deutsche Telekom Wholesale summed up the ambition: "The strategic importance of this Catalyst is that it connects AI-driven automation with trust, accountability, and operational control." The real question an operator needs answered isn't just whether AI can recommend a change, but why it recommended it, which policy it relied on, what risk it flagged, and who or what approved it.

The targets are concrete. The Catalyst aims to cut the time needed to turn a new regulatory or contractual requirement into enforced network access policy from weeks down to hours, and to reduce IAM integration effort by replacing per-system access logic with a single DIM enforcement point. It also targets near-real-time auditability: answering "who accessed what, when, and under what authority" without days of cross-system log stitching.

That last point opens doors beyond day-to-day operations. Sovereign, defence and critical-infrastructure contracts increasingly require contractually enforceable, auditable access control, and static, role-based access models will struggle to qualify for that class of business.

Tallence's role centered on IAM engineering depth. As Marc Seidemann, Tallence's Chief Data Officer, explained: "Trust is not a feature you add at the end. It has to be the foundation from which everything else is built." The Digital Identity Management layer at the heart of the Catalyst isn't a wrapper around existing tooling; it's a principled enforcement architecture in which every access decision is grounded in verifiable policy, every change is bounded, and the audit trail is part of the system rather than an afterthought.

That view comes from experience. Tallence has spent more than two decades designing and operating large-scale identity infrastructure for Europe's leading telcos, including an Identity Management System handling over 50 million accounts and more than 30 million requests a day. As Seidemann put it: "Identity is the primary data problem in network operations. Who is allowed to do what, based on which policy, under which contractual obligation? Without clean identity data, no AI agent can make trustworthy decisions." An AI Compliance Agent is only as reliable as the identity data it reasons over, and getting that foundation right is exactly the discipline Tallence brings to IAM engagements.

Trusted Agentic AI for Access Management is a signal of where enterprise telecom is heading: compliance and access control are becoming live, automated, machine-checkable properties of the network itself, not periodic paperwork. For operators weighing their own path into agentic operations, the questions this Catalyst answers, who is acting, under what authority, and can that be proven after the fact, apply to every system in their identity estate, not just this one project.

Tallence has spent years building the identity foundations that make that kind of trust possible at scale. We're glad to have helped prove out what comes next, and we're ready to keep building it with operators who want to move fast without losing control.

Want to talk about trusted, auditable access management for your network, human or agentic? Get in touch.