The promise of agentic AI is compelling, but for most enterprises, the path to realizing it runs directly through the quality of their integration foundations. Without the right architecture underneath them, AI agents can’t access the data they need, can’t be governed at scale, and can’t deliver the autonomous value that business leaders are expecting.
At Connect:AI London, technology pioneers from Capita, AAH Pharmaceuticals, and GXO shared how MuleSoft has become the connective tissue that is making their agentic ambitions a reality. Their stories offer a practical blueprint for any organization navigating this transition.
MuleSoft as the operational nervous system
For large, complex enterprises, integration is a business-critical capability. When systems fail to communicate, orders stall, customer experiences break down, and AI agents cannot function. What Capita, AAH Pharmaceuticals, and GXO have in common is a recognition that MuleSoft isn’t just an integration tool, but the foundation that connects people, data, and now agents across their entire organizations.
Each has approached the challenge differently, shaped by their industry, scale, and strategic ambitions, but all point to the same conclusion: companies that will “win” in the agentic era are the ones that treat integration as a strategic priority, not an afterthought. Their experiences at Connect:AI London offer a clear picture of what that looks like in practice.
Breaking down silos and redefining relationships with data
Collaboration and data interoperability are significant challenges for large organizations, particularly those growing through acquisition or operating in highly regulated sectors. Without effective, governed communication channels between systems, organisations risk a siloed mentality that slows everything down from day-to-day operations to strategic AI rollouts.
Standardizing integration at GXO
For GXO, a global logistics leader, integration is the business itself:
“If integration fails, orders are not flowing in, goods don’t move. It’s business-critical.”
– Mario Parmentola, Global Integration Leader at GXO
By adopting a strategic approach with MuleSoft, GXO has replaced fragmented point-to-point integrations with reusable patterns that accelerate customer onboarding and reduce the cost of connecting new partners and systems. The discipline that underpins GXO’s integration estate is now forming the blueprint for how the business will scale its AI agents, safely and under control.
Governing agentic AI at Capita
At Capita, a UK business process outsourcer with 34,000 employees, agentic AI is central to the company’s future. Its ambition is to become the UK’s leading AI-enabled business process provider by transforming how the public and private sectors deliver services using data and technology. To support that ambition, Capita moved quickly to establish centralized governance for its agent ecosystem.
“We jumped straight into MuleSoft Agent Fabric when it launched for agent registry and governance. It’s going to really accelerate our identity process automation.”
– Tiina Stephens, Director of AI Engineering at Capita
Stephens also notes that Capita anticipated rapid growth in the number of agents it would need to manage: “We recognized that AI agents are likely to grow in the same way as apps. You make a best estimate and then triple that number. MuleSoft has given the answer to that problem for us because we can now apply enterprise-level AI governance at scale.”
Frictionless customer and supplier journeys
The business case for autonomous AI agents via improved resource allocation, faster query resolution, and 24/7 service availability is becoming tangible for organizations that have built the right foundations. Two of the Connect:AI London panelists are already seeing measurable returns.
AAH Pharmaceuticals’ virtuous cycle
AAH Pharmaceuticals, the UK’s largest drug wholesaler, is eight years into its MuleSoft journey, and the compounding benefits are clear.
Ranjit Gill, CIO of AAH Pharmaceuticals, describes a “virtuous cycle” in which frictionless digital transactions are making it cheaper and easier to deal with customers while simultaneously improving the company’s reputation for responsiveness and efficiency. Today, nearly 100% of customer interactions are digital, governed by a suite of APIs that manage the secure flow of information across the business. Revenues from digital channels are up 70%.
“As digital platforms continue to evolve into agentic-driven commerce, using MuleSoft with a headless approach will allow us to navigate the next few years. By handing repeatable tasks to AI agents, we release people to do higher-value work.”
– Ranjit Gill, CIO of AAH Pharmaceuticals
He’s clear that governance is what makes this possible: “Value is already being generated by autonomous agents. It’s about having the governance controls and standards to allow focused agentic functions to deliver value across the business.”
Measuring success at GXO
At GXO, the metric that best captures the business impact of MuleSoft is time to onboard. As Parmentola explains: “It’s simple but powerful. If onboarding time drops, it means we’re reusing more, spending less, and standardizing effectively.” That focus on reuse and standardization is what allows GXO to scale its integration estate without a corresponding increase in complexity or cost – a principle that will apply equally as the business extends the same discipline to its AI agent deployments.
Centralized governance with regional agility
One of the most important lessons from all three organizations is that governance and innovation are mutually reinforcing one another. The key is building a governance model that sets clear enterprise-wide standards while giving regional and business-unit teams the freedom to execute within those guardrails.
Capita recognized early that it needed a centralized hub for agentic orchestration. Rather than allowing each business unit to manage its own AI platform independently, Capita uses MuleSoft Agent Fabric to ensure a single, consistent vision of governed AI across the organization – preventing the fragmented landscape of shadow AI that can quickly undermine security and compliance.
Stephens notes: “We wanted to bring a democratized approach to AI agents, but we need to make sure we’re governing that in the right way whilst not preventing innovation.” The results speak for themselves: Capita has scaled to more than 400 agents in six months across a multi-platform, multi-cloud environment.
At GXO, the MuleSoft Center of Enablement has played a similarly pivotal role. By establishing global standards for platform stability while allowing regional teams to execute within defined guardrails, it ensures that innovation at the edges of the organization doesn’t compromise the integrity of the core.
As Parmentola summarizes: “The discipline we’re bringing within the integration space in terms of standards and processes is forming the foundations for the general business. The same principles that let us scale our integration safely will let us scale our AI agents in a secure, controlled way.”
Integration is the foundation for agentic success
Capita, AAH Pharmaceuticals, and GXO’s experiences make one thing clear: the path to agentic AI success isn’t purely about agent intelligence, it’s about the integration and governance that allow that intelligence to act safely, reliably, and at scale on behalf of your business.
MuleSoft provides the system that connects people, data, and agents, turning isolated AI initiatives into a coherent, governed, and high-performing agentic enterprise. For organizations still in the early stages of this journey, the message from these technology leaders is simple: if you start with a strong foundation, the possibilities are infinite.
