Capita has launched the Forward Deployed Orchestrator™ (FDO), a new operating model that embeds an AI-fluent process expert inside a client's operation to own the results of agentic AI after go-live. The move, announced on 1 July 2026, addresses a persistent headache for operations leaders: AI tools can be built quickly, but making them work safely, reliably and at scale in regulated back and middle offices is far harder.
Many organisations have struggled to move AI beyond pilots. MIT's 2025 study of enterprise generative AI found that roughly 95% of organisations see no measurable business impact from the technology. McKinsey has reported that fewer than one in ten have scaled AI agents in any single function. The hard part starts after deployment, when tools need to be embedded in live workflows, governed properly and continuously improved as rules and customer needs change. Without clear operational ownership, AI investments rarely deliver lasting value.
Where the accountability gap lives
"All of our clients are under pressure to deploy AI, but far fewer have anyone accountable for whether the AI continues to work once live," said Sameer Vuyyuru, Chief AI and Technology Officer at Capita. "Building an agent is now the easy part; the real challenge starts in live, regulated middle and back-office operations, where a wrong answer carries real consequences."
The FDO is designed to step into that gap. Rather than building and walking away, the orchestrator runs the AI-enabled process in the live environment, drives adoption so the new way becomes the standard way, and tunes performance across a combined human-and-machine workforce. The role is platform-agnostic, accountable for a measurable outcome and focused on continuous improvement - not just a successful initial deployment.
Tested on Capita's own operations first
Capita proved the model on its own high-volume recruitment process, using Salesforce's Agentforce technology. An internal FDO pod worked alongside recruitment, people and operations teams as "customer zero." In the first four months, the approach saved more than 1,000 management and recruitment hours and cut candidate screening time by 43%.
In another regulated service, straightforward applications that previously averaged four days to clear are now completed in about 11 seconds. The AI reads and verifies documents, runs authenticity and fraud checks, and refers only cases needing human judgment to a colleague with the right expertise. More detail on the model is available on Capita's Forward Deployed Orchestrator page.
Why this matters for Operations
The FDO model gives operations executives a straightforward fix for the accountability vacuum that kills many AI initiatives. Instead of treating AI as a one-off project, it embeds a person who owns the live process and its ongoing performance - critical in regulated environments where auditability and human oversight are mandatory. The role demands process expertise, change management and AI fluency, skills that dedicated programmes for AI for Operations Managers aim to build.
Your membership also unlocks:
