Thank you for subscribing to the WSO2 Newsletter

Newsletter - August 2026

Thank you for subscribing!

Take a look at our latest newsletter and watch your inbox for the next one.

IN THE NEWS

A new CEO, and a team built for WSO2's next stage of growth

WSO2 is putting new leadership behind our vision. Harry Ault has joined as Chief Executive Officer, following a year that already added a new Chief Financial Officer and Chief Marketing Officer to the executive team, all in support of WSO2's next stage of growth. The company that has spent 20+ years building open source infrastructure for the world's enterprises now has the team in place to carry Trusted AI Governance to a global customer base built on an open, composable, sovereign foundation.

WSO2 joins the Agentic AI Foundation

WSO2 has joined the Agentic AI Foundation (AAIF), reinforcing our commitment to an open agentic ecosystem built on shared standards rather than proprietary technology. The Linux Foundation initiative stewards key open standards including MCP and A2A, and builds on WSO2's existing collaboration with the foundation through summit participation, ambassadorship, and contributions to standards papers. As a member, WSO2 will continue helping shape the standards that agents and agentic systems rely on, because AI governance starts with an open foundation.

Sovereignty with WSO2

WSO2 partners with governments and enterprises to help them implement sovereignty over the systems that govern their identity, data, and increasingly, the AI agents acting on their behalf, regardless of where they deploy them or which infrastructure and vendors they choose. That control starts with open, auditable software they can inspect, modify, and run on their own terms, without being locked into proprietary technology. And it’s already proven at scale: WSO2 technology is already used by 42 national governments and more than 3,800 local government agencies worldwide. Now, the same principle of control is extending to AI agents, giving organizations the ability to govern what they can access, what they can do, and who they act for.

Your enterprise architecture just got a new kind of user

AI agents are increasingly acting as first-class users of enterprise systems, calling APIs, chaining workflows, and executing actions without ever touching a UI. In a recent Techzine blog on AI agents as users of enterprise architecture, WSO2's Chief AI Officer Rania Khalaf unpacks three shifts technical leaders need to plan for: API surfaces replacing dashboards as the real user experience, the need for "agent-visible" product capabilities, and how non-deterministic agent behavior is upending traditional testing, observability, and guardrails.

Why your APIs aren't ready for AI agents yet

APIs were designed for deterministic, human-driven traffic, but AI agents call them in exploratory, non-linear, and probabilistic ways. Nuwan Dias, Vice President and Deputy Chief Technology Officer at WSO2, breaks down why MCP alone doesn't solve this problem in a blog on API design for AI agents, and can even amplify weaknesses in APIs that aren't AI-ready, outlining practical fixes like intent-oriented operations, machine-readable error codes, idempotency, dry-run support, bulk endpoints, and AI-ready capability discovery.

What does it take to trust an AI agent in production

AI coding tools work well on a laptop, but running agents against production infrastructure raises harder questions around abstraction, credentials, and trust. In a recent conversation, OpenChoreo maintainers Lakmal Warusawithana and Sameera Jayasoma explore what it takes to make agents viable in production; from giving them higher-level interfaces to treating them as scoped, auditable platform users rather than privileged actors.

What happens when an AI agent gets it wrong?

AI agents can execute complex workflows, but when something goes wrong, they can leave behind unintended side effects and may not reliably remember everything they've already done. In a paper presented at ACM CAIS 2026, WSO2 SVP and Head of Research, Srinath Perera introduces Robust Agent Compensation (RAC), a safety net that keeps a recovery log of an agent's actions and uses it to retry, find alternatives, or compensate for failed operations. Across three benchmarks, RAC outperformed the approaches it was tested against, including getting 8 of 9 previously unseen problem scenarios right. It's a different approach to making agent execution predictable.

Traditional IAM wasn't built for agents that act on their own

AI agents break the foundational assumptions traditional IAM was built on: identities that are stable, discrete, and human-anchored. Ayesha Dissanayaka, Associate Director/Architect at WSO2, examines how the four pillars of IAM, Administration, Authentication, Authorization, and Audit, must evolve to govern agents that act simultaneously as subjects, resources, and clients, principles that WSO2 Agent ID puts into practice on open standards like OAuth 2.0, OIDC, and MCP.

Inside WSO2's first Africa-based office, and what's driving it

WSO2 has opened its first Africa-based office, in Johannesburg, South Africa, backing a region that has delivered roughly 30% year-on-year growth and expanded from around 40 to more than 65 customers across 25 countries, including public-sector organizations like Uganda's National Information Technology Authority and the Kenya Revenue Authority. Dinal Gooneratne, WSO2's Vice President and General Manager for Africa, explains what's driving that growth and where the region is headed next in a recent interview with IT News Africa.

See it all come together in Nairobi

WSO2Con Africa 2026 is happening September 22 to 24 at the JW Marriott Nairobi, and we'd love to have you there. Expect sessions on the strategy, architecture, and technology behind Trusted AI Governance, hands-on time with WSO2 Agentic Enterprise Fabric, and direct access to the engineers, architects, and leadership building it. View the agenda to see who's speaking.

CTO’S PERSPECTIVE