Client Profile

Customer Name: Barts Health NHS Trust

Established: 2012

Customer Since: 2005

Employees: 16900

Sector: Healthcare

Barts NHS Trust is one of the largest healthcare providers in the UK, serving millions of patients across east London and beyond. The trust operates five hospitals, including the Royal London Hospital, a major trauma centre and tertiary teaching hospital. Every day, thousands of clinical and non-clinical staff work to deliver emergency care, specialist services and routine procedures to the communities they serve. 

Like all NHS trusts, Barts manages complex financial operations across procurement, invoicing and accounts payable. Processing supplier invoices accurately and efficiently is critical, it underpins everything from paying staff to ordering medical supplies. 

The opportunity 

Barts processes approximately 400,000 supplier invoice pages every year. Their existing approach relied on template-based rules to extract invoice data. As supplier formats evolved and business needs changed, the trust recognised an opportunity: could modern AI and cloud technology transform this process? 

Barts wasn’t responding to a crisis, the existing system processed invoices, and it worked. But the trust’s leadership saw something the numbers didn’t show: NHS organisations that wait for a system to break before changing it tend to be behind. They wanted to know whether AI could do something more useful than follow rules, whether it could handle the messiness of real supplier data without a human rewriting the template every time a new format arrived. That’s a harder question to answer than “can we fix this”. 

Four people around a desk

Why Version 1 and Oracle OCI 

Barts chose Version 1 and Oracle Cloud Infrastructure because the partnership offered something specific: rigorous, enterprise-grade AI built on proven cloud infrastructure. 

The technology stack matters here; Oracle Document Understanding AI isn’t a generic tool. It’s purpose built for complex business documents and learns from patterns. Combined with Oracle Integration Cloud for orchestration and Oracle APEX for modern user interfaces, it represents a fundamentally different approach to document processing. 

Barts saw this as a genuine innovation opportunity. Not a vendor replacement project, a transformation project. 

What’s being built 

Version 1 is delivering an intelligent document processing solution built on Oracle Cloud Infrastructure. At its core, OCI Document Understanding AI handles extraction, using pre-trained models as default, with generative models applied selectively for complex supplier formats. Oracle Integration Cloud manages routing and validation, Oracle APEX provides the staff-facing dashboard, and everything connects directly into Oracle EBS where Barts’ financial records live. 

The operational model is designed around confidence scoring. Every extraction gets a score per field. High-confidence invoices move faster through review. Lower-confidence invoices surface to staff with context already assembled, so the decision is faster and better informed. The system doesn’t remove human judgment, it reserves it for the cases that actually need it. 

Version 1 is also building in the change management work that technology projects often skip. Healthcare organisations have ingrained ways of working. Phased implementation, early staff involvement, and clear governance around how the system learns and makes decisions are part of the delivery, not an afterthought. 

Healthcare-sector-image-of-person-in-science-lab

What this reveals about AI in healthcare 

Six months into delivery, a pattern is emerging; AI in healthcare isn’t about replacing people. It’s about amplifying their judgment. 

Finance teams are not going anywhere. What changes is how they spend their time; less keying in data, more deciding what to do when something unusual lands. The routine work gets handled, and the judgment calls still need people. 

It’s why Barts sees this as a people-first transformation, not a technology play. 

The trust is acutely aware that AI at scale in healthcare requires rigour. This is foundational to the NHS, and requires clear data governance, explainability, audit trails, and the understanding of why the system made a decision.  

“Version 1 understands healthcare complexity and they take a rigorous approach to AI implementation. This isn’t a quick technology fix for us, it’s a genuine transformation of how we process invoices and how our teams work. We’re pioneering something here, and that requires a partner who understands both the technology and the organisational change required. When we go live, I think we’ll have something genuinely innovative to show the NHS.”

Jason Britchford, Barts NHS Trust Finance Leadership

What comes next 

Barts is tracking towards go-live in the coming months. Once live, the real learning begins: how does the AI perform with real-world supplier data? How do teams adapt? What refinements emerge? 

Version 1 will be monitoring closely. Not just whether the system works, but whether it genuinely transforms how Barts’ finance teams work. Whether staff engagement increases. Whether it becomes a repeatable model for other NHS trusts. 

The stakes are high. Not just for Barts, but for AI adoption across the NHS. If intelligent automation can work in healthcare at scale, it opens possibilities far beyond invoice processing. 

 

For more information on how Version 1 can help transform public health care through AI and data technology that works, get it touch: Transforming Public Healthcare with AI & Data | Version 1