In this episode of the CloudTweaks Podcast, host Steve Prentice sits down with Francis Martens, CEO and co‑founder of Exalate, to explore one of the most urgent and least understood challenges in modern enterprise technology: integration in the age of AI agents. As Steve frames it in the opening, AI agents are beginning to act across systems, teams, departments, and company boundaries at machine speed, and the long‑ignored integration layer connecting these environments is suddenly the new frontier of risk, complexity, and opportunity.
Francis explains that traditional integration tools were built for internal workflows: single organizations with unified governance, shared semantics, and predictable change cycles. But once you cross into multi‑system or multi‑company ecosystems, everything changes. “Whenever you integrate your own environment with five or six different other companies,” he notes, “you have no control on how changes are being applied on the other side.”
This is where Exalate’s distributed, loosely coupled architecture stands apart. Designed to maintain data consistency, workflow independence, and resilience to change, Exalate enables organizations to synchronize systems like Jira and ServiceNow without forcing either side to conform to the other’s internal structures. And its value extends beyond cross‑company use cases: Exalate connects disconnected systems across teams, tools, and departments. Customers even use “Exalate” as a verb, reflecting its growing influence in the integration landscape.
AI agents are accelerating the problem and the need for better solutions The rise of agentic AI is amplifying both the complexity and urgency of integration. AI‑generated configurations can be powerful but unpredictable. Francis warns that AI can create integrations that are far more complex and difficult to maintain, and that organizations will soon rely on agents to perform actions across external systems at scale. This creates a paradox: AI makes integrations easier to build, but exponentially harder to govern.
Governance, trust, and the human‑in‑the‑loop
Governance has always been essential in cross‑system integration, but AI makes the stakes higher. Francis emphasizes that AI doesn’t create the need for governance, but it certainly accelerates it. Exalate becomes not a reaction to AI hype, but the layer that was always necessary and is now impossible to delay.
He uses two vivid analogies from the automotive world. First, with self‑driving cars: when an autonomous agent acts across systems and something goes wrong, who owns the decision and the outcome? This question now extends from IT into the boardroom. Second, he imagines the chaos that would come from AI regulating traffic lights and trying to please all sides at once. The same applies within organizations: governance must be built in before AI acts, not after.
Exalate’s approach is clear:
• AI may help configure integrations
• But the execution engine must remain predictable, controlled, and trustworthy
• And a human‑in‑the‑loop is essential to validate guardrails before deployment
iPaaS as plumbing and why it’s no longer enough
Francis describes traditional iPaaS platforms as plumbing: useful for moving data, but not designed for the dynamic, multi‑party, AI‑driven workflows emerging today. AI agents may still use iPaaS as a conduit, but the intelligence and governance layers must evolve beyond what iPaaS was built to handle. A turning point: “The iPhone moment of 2007”
Francis compares today’s AI‑integration landscape to the dawn of the smartphone era: transformative, poorly understood, and full of unknowns. The only way forward is experimentation, transparency, and shared learning across the industry.