The conference further spotlighted the road to autonomy, outlining the steps operators must take to evolve from reactive approaches to augmented, agentic and ultimately states.

A standout session came from a panel featuring Harsha Angeri, VP corporate strategy and head AI business at Subex, joined by Mohamed Temsamani, director of revenue secure at Inwi. Their discussion offered a view into how operators are confronting a rapidly evolving fraud landscape, and how strategic partnerships are reshaping industry readiness.

Angeri opened with a reflection on the pace of AI development and its implications for telcos. With Google’s latest model release accelerating global capability, he stressed that fraud management teams must match that speed.

“The competition is changing, the game is changing, and you have to win differently,” he said. For Subex, this means shifting customers from manual and rule-based systems to adaptive, agent-driven fraud management that can scale autonomously as threats evolve.

Setting the broader context, Angeri emphasised the economic urgency. Fraud continues to grow at nearly 12% CAGR, eating into margins at a rate far outpacing EBITDA growth for most operators.

Subex’s data suggests that AI-based uplift can deliver 35–40% performance improvement when layered onto rule-based systems, and up to 99% accuracy for net-new use cases. The prize, he argued, is not incremental but transformational: a potential US$60–155 billion in enterprise value unlocked globally.

It was against this backdrop that Temsamani shared first-hand experiences of deploying Subex’s AI model within their respective organisations.

Inwi: Shifting time from detection to analysis

For Morocco’s Inwi, the imperative was clear. “Our traditional sources are not enough to detect emerging fraud,” Temsamani said. Investigators were spending excessive time searching for issues and too little time analysing new patterns. AI agents offered a way to reverse that dynamic.

Working with Subex, Inwi has already seen a dramatic improvement in investigation efficiency. Angeri referenced internal statistics showing average investigation time (From detection to closure) reduced by more than 85% The solution is operable 24/7, with each case costly only a few cents to resolve.  Temsamani confirmed the impact: “Now we can reduce false positives, reduce case volume, and dedicate more time to real fraud.”

Beyond interconnect and IRSF fraud, Inwi is now exploring how AI agents can scale to new forms of digital and network-based fraud detection and investigation. The partnership model with Subex, a joint engineering and FM business function supported by data science profiles has allowed the operator to accelerate capability building across teams.

According to Temsamani, this separation of engineering and fraud management roles, while enabling close collaboration, has proved essential, with engineering and data science becoming integral to fighting fraud: “We give the engineering team enough time to analyse new patterns and support the FM team. This is the strength of our organisation.”

The road ahead

The panellists agreed that fraud patterns are evolving, particularly as social engineering rises sharply. While many fraud typologies remain structurally similar, attackers now wield more advanced tools, faster iteration cycles and increasingly multi-agent strategies. The recent high-profile cyber incidents referenced by Angeri underscore that adversaries are already using sophisticated agent systems.

For example, Anthropic recently reported multi-agent attacks by state actors using generative AI to target multiple organisations simultaneously.

Looking forward, Inwi expect AI agents to become core operational partners within fraud teams. At Inwi, the fraud management function is increasingly embedded with engineering and data science talent.

RELATED STORIES

Subex CEO Nisha Dutt on rebranding, AI strategy & the future of telecom

How AI is transforming telecom: Subex VP Harsha Angeri on automation & fraud detection

VodafoneZiggo’s Angelo Lemmens on diversification, digitalisation and the next phase of growth