Salesforce has opened a new AI center in London, positioning the site as a hub for customers, partners, public sector organizations and startups to explore agentic AI use cases.
“We’ve moved past pilots,” said Zahra Bahrololoumi CBE, CEO of Salesforce UKI, speaking at the centre’s opening. “We’re now in the era of execution, and execution at scale.”
The move forms part of Salesforce’s five-year, $6BN investment in the U.K., which the company said is focused on AI innovation, customer adoption and workforce development. It also indicates how the company is positioning Agentforce, its agentic AI platform, as enterprises move from experimental use cases into more operational deployments.
The vendor said Agentforce is now deployed by more than 23,000 customers globally, with U.K. organizations including Heathrow Airport, Pets at Home, NHS Shared Business Services, the National Trust and police forces across England among those using the technology.
Speaking during a panel discussion at the opening, Paul O’Sullivan, Chief Technology Officer at Salesforce UK&I, said AI adoption is starting to shift from theoretical opportunity to measurable use cases.
“We’re starting now to really see tangible business or citizen value associated with how AI can help. Technology trapped inside the box means nothing to no one.”
“Technology out there in the hands of the humans that need it, or use it, is where we’re going to see the most amount of value.”
Agentforce Use Cases Move Into Customer and Citizen Service
Salesforce highlighted several use cases that center on service automation, where AI agents are being deployed to handle customer, employee and citizen interactions.
In policing, the AI agent “Bobbi,” built on Agentforce and deployed by Thames Valley Police and Hampshire & Isle of Wight Constabulary, handles more than 200 non-emergency conversations daily and resolves 75 percent of those interactions autonomously, according to Salesforce. Bahrololoumi also cited broader adoption across three UK police forces, saying Bobbi is helping to resolve 82 percent of inbound non-urgent enquiries without escalation to a human officer.
In healthcare, NHS Shared Business Services has used an AI-powered help center built on Agentforce to reduce case resolution times from five days to 24 hours, the vendor said.
Salesforce is positioning agentic AI as a way to help public services absorb routine demand while freeing up employees for more empathetic or judgment-led work.
Bahrololoumi said:
“Rapid adoption of our AI agents is supporting stretched public services and giving organizations across the U.K. the capacity to deliver more.”
NHS Shared Business Services has reduced its case resolution times from five days to 24 hours using an AI-powered help center built on Agentforce.
These deployments illustrate Salesforce’s argument that agentic AI is moving into core service workflows.
Salesforce Points to Its Own AI Center as an Internal Test Case
Salesforce has increasingly used the “customer zero” message around its AI strategy, emphasizing that it is deploying Agentforce and related AI capabilities within its own operations before encouraging customers to adopt similar models.
O’Sullivan described the AI center’s resident team as an example of how the company is using AI internally to increase output without necessarily scaling headcount at the same rate.
“If we think about the resident team in the AI center, it’s eight people. Those eight [engineers], augmented with AI, have evolved to learn how to orchestrate AI in a way in which they can get more value and throughput and get more done.”
Still, customers are likely to judge internal productivity examples against more practical questions, including where AI can safely remove friction, where human oversight is required and how organizations can measure the impact beyond activity or output.
O’Sullivan argued that this requires AI to become part of everyday operating models rather than being treated as a separate innovation track. “When we think about humans at the helm, this is where I think it has to be embedded into everything we do.”
O’Sullivan added that organizations should avoid simply layering AI onto legacy processes.
“You have to reimagine your processes. Don’t just apply AI to the same [things] you’ve always done. Think about: is this an opportunity to reimagine all of those processes—how you operate and how you deliver?”
The AI Center’s Role in Customer Adoption
The new location builds on Salesforce’s first London AI center, which it opened in 2024, but relocates and expands that concept. The facility includes a learning lab, collaboration zones and an immersive AI theater, with hands-on demonstrations showing AI agents solving business challenges. Salesforce said it expects to host more than 5,000 visitors this year.
The company intends the center to function as a demonstration space and a workshop environment where organizations can identify use cases and explore how agentic AI might fit into their operations.
When working with customers, Salesforce typically starts by observing how employees currently use technology, O’Sullivan said. The next stage is to encourage broader thinking about what could change.
“We then typically take our customers through a process of playback. We’ll talk them through our observations, and then what we try and encourage is the concept of divergent and convergent thinking.”
That approach is key for CX technology adoption because many AI opportunities sit in the handoffs between systems, teams and customer-facing processes rather than isolated tasks.
For example, a service agent moving between multiple systems to answer a customer query may appear to be dealing with a technology issue. But the larger problem may be fragmented knowledge, unclear process ownership or disconnected data.
By focusing on those workflow gaps, Salesforce is positioning Agentforce as part of a broader process redesign conversation.
Relina Bulchandani, Executive Vice President of Real Estate and Workplace Services at Salesforce, said that the company had made a deliberate decision to create a physical space for AI education and collaboration, despite the increasingly digital nature of work.
“We designed this campus to be a catalyst for innovation, showing what becomes possible when humans and AI agents work together.”
The use cases that Salesforce highlighted indicate that the vendor is increasingly positioning Agentforce as a live service platform rather than a future-facing AI experiment. The examples from policing and NHS Shared Business Services show Salesforce leaning into measurable service outcomes, including containment rates, reduced resolution times and increased capacity.
The center will also be available to Salesforce Ventures portfolio companies and the broader U.K. AI ecosystem. Salesforce Ventures has invested more than $250MN in U.K.-based AI companies, including AutoGenAI, Covecta, and ElevenLabs.
The company is using the campus to accelerate enterprise confidence in AI agents. By giving customers and partners a physical environment to test and workshop AI use cases, the vendor is trying to shorten the path from curiosity around AI to operational deployment.
