Why Cequence Is Co-Chairing TM Forum’s AI-Native Blueprint Initiative
This past Sunday, at Mobile World Congress in Barcelona, TM Forum officially named Cequence Security as Co-Chair of its AI-Native Blueprint Initiative, specifically leading the Agentic Interaction Security workstream. For those of us who spend our days thinking about application and data security, and the emerging attack surface created by AI agents, this is more than just a title; it’s a call to action for the entire telecommunications industry.
The Problem Is Real, and It’s Moving Fast
Let’s start with the numbers, because they tell an uncomfortable story.
More than 80% of Fortune 500 organizations now deploy active AI agents, yet only 47% have AI-specific security safeguards in place (Microsoft Cyber Pulse, February 2026). Only 14.4% of AI agents launch with full security approval. And 48% of security professionals rank agentic AI as the number one attack vector for 2026.
Meanwhile, MCP vulnerabilities grew 270% from Q2 to Q3 2025 alone, with Coalition for Secure AI (CoSAI) identifying a dozen distinct MCP threat categories in their Jan 2026 whitepaper. This isn’t a future problem. It’s happening now, at scale, in production environments.
For telecommunications providers, the stakes are especially high. Telcos are deploying AI agents across internal productivity tools, customer experience platforms, and autonomous network operations, all while sitting on some of the most sensitive data in existence, including subscriber PII, payment data, IMEI, and CPNI. Every agent interaction with those environments is a potential path to the crown jewels.
Why the Telecom Industry Can’t Go It Alone
The core challenge is one of asymmetry. Threat actors are already exploiting the gaps – rogue MCP servers can coerce trusted agents into exfiltrating sensitive data or executing unauthorized actions, hallucinations can trigger unintended modifications to critical business systems, and tool poisoning can compromise entire agent workflows. The industry, by contrast, is still developing the playbooks.
That’s precisely why TM Forum’s collaborative model matters. No single vendor, no matter how capable, can solve an industry-wide security problem by itself. The AI-Native Blueprint Initiative brings together leading global CSPs and technology partners to build interoperable, governed, and trusted agentic AI frameworks, which are the kind of shared security baseline every operator needs but none can build alone.
What Cequence Brings to the Table
Cequence isn’t coming to this initiative with whitepapers and theories. We’re bringing a decade of production-grade operational experience protecting the applications and data of some of the world’s largest telecoms.
Consider what that looks like in practice: at one top-three U.S. carrier, we’ve discovered tens of thousands of previously unknown application interfaces and mitigate more than 1 billion malicious or abusive requests every week. At another, we secured an environment spanning more than 18,000 interfaces and identified 115 high-severity issues, including numerous RCE vulnerabilities. Across our telecom customer base, we analyze roughly 60 billion application interactions per month, 37% of which are classified as malicious or abusive.
In total, Cequence protects more than 10 billion sensitive interactions per day and more than 4 billion user accounts worldwide. Two of the top three U.S. carriers are customers. That’s the source of the operational knowledge and experience we’re now bringing to the broader industry through TM Forum.
As Co-Chair, our focus will be on contributing real-world threat intelligence to help define frameworks for secure agent interactions, integrity standards for agentic AI connections including the Model Context Protocol (MCP), real-time governance blueprints, and operational resilience patterns that organizations can actually implement.
This Is Bigger Than Telecom
While this initiative is anchored in the telecommunications sector, the security patterns we develop will matter to any enterprise deploying agentic AI at scale. The intersection of application and data protection, business logic abuse, and agentic AI is not being addressed anywhere else with the depth of operational experience that telecom-scale environments demand.
Cequence is also the only security vendor in its category contributing to three consecutive Verizon DBIRs, and our production-proven AI Gateway technology is already securing MCP-powered access to applications and data in live environments across verticals. Further, this TM Forum work complements our contributions to emerging CIS guidance for agentic AI and MCP environments, all part of a deliberate philosophy of giving back rather than hoarding knowledge for competitive advantage.
The Bottom Line for Security, IT, and AI Leaders
If you’re responsible for securing and scaling agentic AI deployments that work, whether you’re in a telecom environment or not, the patterns and guardrails that come out of this initiative will be directly relevant to your work. The agent economy is scaling whether your security posture is ready or not. The question is whether the industry builds the shared frameworks to manage that risk before threat actors fully exploit the gap.
We’re committed to making sure the answer is yes. And we’re glad to be doing it alongside TM Forum and the global telecommunications community.
Read the full press release here.
