BLOG | Trust is not a feature. It's the foundation.

At DTW Ignite 2026 in Copenhagen, a single finding set the tone for the whole event. In his opening keynote, TM Forum CEO Nik Willetts revealed that while 72% of CSPs say their AI is trustworthy, only 14% can produce evidence to prove it. "Get trust right," he said, "anything is possible. Get trust wrong, and nothing else matters."

That research landed as a wake-up call. And it framed the conversation at the close of day two between Guy Lupo, EVP, Trustworthy AI and Data at TM Forum, and Arnab Chakraborty, Chief Responsible AI Officer at Accenture and recently appointed to the TM Forum Trustworthy AI and Data Mission board.

Their conversation covered a fundamental industry challenge in the age of the AI-native telco: how do you operationalize trust?

From principle to practice

What makes Arnab's perspective distinctive is not just his role at Accenture. He is also joint CEO of the Accenture-Telstra Data and AI Joint Venture, which means he isn't advising on trust from a distance. He is living it, as both service provider and, in effect, operator. That combination puts the real-world challenges in sharp relief.

The opportunity, as Arnab sees it, is enormous. The telecom industry stands to benefit from $60 to $100 billion of economic value annually from AI through elevated customer experience, enterprise autonomy, and the journey to level 4 autonomous networks. "These are path-breaking opportunities," he said. But the gating factor is trust.

The question he keeps returning to is this: how do you ensure that every stakeholder — customers, employees, shareholders, regulators — can trust AI at every step of the lifecycle? "If you can actually build that trust, then there will be adoption, and scale, and the value realization will happen."

And there's a further opportunity for operators who get this right. A truly trustworthy operator doesn't just run better networks, it earns the right to build AI-based products and services that go well beyond connectivity. Trust, done right, becomes a commercial differentiator.

The four Ps: making trust real

Guy's question cut to the heart of where most of the industry is stuck: knowing what needs to be done is one thing. Knowing how to operationalize it is another.

Arnab's answer was concrete. He laid out what he calls the four Ps of operationalizing trustworthy AI at scale.

The first is policy — AI policy that is embedded by design into every AI solution, not added as an afterthought. The second is process — embedding that policy into actual practice across the full AI lifecycle, from design through build, deployment, post-deployment, and all the way to retirement. Trust has to be present at every stage.

The third is platform — a tooling chain that creates and captures the evidence of trust being built across the entire value chain. This is where the sceptics, the security teams, and the regulators will look. And the fourth, equally critical, is people — equipping everyone in the organization with the curriculum and understanding to use AI systems in a responsible and trustworthy way. Arnab has seen the impact of this firsthand across Accenture, and at a leading telco like Telstra.

The control plane and the chain of evidence

On the question of proof — audit trails, evidence, the things regulators actually need — Arnab was direct. The answer lies in an AI control plane: a layer through which AI systems are scaled, and which captures every decision made, every data source accessed, every instance of drift. Guard rails create alert mechanisms. Everything gets captured in a digital trail. When a regulator arrives, you’re able to present a complete record from design to deployment.

"You are not just reacting," Arnab said. "You are actually shaping the agenda of what it looks like when it comes to trustworthy AI."

Building the North Star

As a new member of the Trustworthy AI and Data Mission board, Arnab's vision for what TM Forum can achieve is ambitious. It is also deliberately industry-wide. The goal is to build trustworthy AI frameworks by the industry, for the industry, and democratize them to every member. Not as a telco-specific project, but as a model that other sectors — healthcare, banking, beyond — can look to and follow.

Telecom, he argued, is the right place to prove this out. It is one of the most heavily regulated industries in the world. Getting trust right here, under that level of scrutiny and complexity, is proof it can be done anywhere.

As Guy closed the conversation, he asked Arnab for one sentence to sum up DTW Ignite 2026 and the journey ahead. The answer was clear: "Trust is paramount. Trust is about creating adoption with AI so that you can realize the autonomous vision. And that's when you actually realize the value."

The Race to 2030 has a clear condition of entry. Build the trust first.

Watch the video at https://inform.tmforum.org/videos/trust-is-not-a-feature-its-the-foundation