Lloyds Banking Group scales GenAI with control

The Business Issue

For Lloyds Banking Group, commercial banking is big by design. The group serves clients across business banking, commercial banking, corporate and institutional banking, insurance, pensions, and investments.

In commercial banking operations, that means supporting a wide range of needs – from everyday account servicing, mandates, cards, and online support to specialist lending, trade finance, asset finance, bereavements, and complaints.

That breadth gives Lloyds Banking Group enormous reach. It also means small points of friction can grow fast. Lloyds Banking Group’s commercial banking operation creates more than two million cases each year, supported by 5,000–7,000 inbound emails every day and roughly 2,000 outbound communications. Much of that work still asked employees to read, interpret, classify, copy, paste, and draft across emails, documents, forms, and case histories. At that scale, a few minutes lost on each case becomes a serious drag on the people doing the work – and on the customers waiting for answers.

Lloyds Banking Group had already made big moves to simplify how work moved through the bank. Commercial banking teams had shifted from a more varied, site-based model to specialist teams, standardized processes, skills-based routing, central email indexing, and single-actor fulfillment. Those changes gave the bank a stronger base for growth. They also uncovered the next challenge: as volumes grew, faster service couldn’t depend on adding more operational effort.

Commercial banking also isn’t work you can leave to self-service and walk away from. Many journeys still need judgment, context, and care – especially complaints, bereavements, specialist lending, and complex client servicing. Lloyds Banking Group needed a way to give employees time back without losing control of the process.

The Solution

Lloyds Banking Group started where the work was hardest on employees. The team looked at the places where people spent time reading long emails, extracting data from forms, reviewing case history, drafting communications, and routing work to the right team. They saw an opportunity to use GenAI to remove some of that manual effort and give them more time back for the work that needed human judgment.

One opportunity stood out: asset finance. The process included 12 process steps, 70 checklist questions, limited data and processing insight, and manual re-keying from PDFs into case forms. Lloyds Banking Group redesigned the Pega case with a structured workflow, automated communications, a full management information suite, and automated data sourcing. Then the team used Pega GenAI Connect to extract key fields from structured PDFs and place them directly into the case.

The team also built the control model before scaling the technology. Architecture, risk, and compliance teams reviewed the approach. Human-in-the-loop checks stayed in place. The team used masking, filtering, logging, and prompt controls. Employees could also give feedback on generated content, including whether an output was helpful or had an issue that needed review.

That control model became the bridge from one use case to many. In that process, employees often had to review lengthy email chains, documents, submitted notes, and case history before they could move forward. Lloyds Banking Group redesigned the case for direct creation, pre-populated data, automated routing, and end-to-end tracking in Pega. GenAI now supports instant complaint summaries, related case summaries, and draft response letter content – with employees reviewing and approving the output before it reaches a client.

That first use case gave Lloyds Banking Group more than a working feature. It gave the team a repeatable pattern for bringing GenAI into commercial banking operations safely. With the first path proven, they moved into more complex journeys, including commercial complaints.

The Results

The asset finance use case gave Lloyds Banking Group clear proof that GenAI could create value inside real commercial banking work. The team has created over 5,000 asset finance cases, saved over 500 hours of manual effort, and reached 99% accuracy in data extraction. The employee feedback shows what those numbers mean in practice:

• “The whole experience feels less clunky and far more streamlined.”
• “No more time wasted copying and pasting – emails sent instantly and processes completed in a flash!”
• “So much easier and quicker. Less back-and-forth and far more straightforward.”

The complaints work points to the next level of impact. The previous process included multiple handoffs and re-keying, lengthy email chains and document review, 160 minutes of average case handling time, and a typical end-to-end resolution time of two weeks. With the redesigned Pega case and GenAI overlay, Lloyds Banking Group expects to save a projected 30 minutes per complex case assessment and a projected 45 minutes in response letter drafting.

For Lloyds Banking Group, the bigger story is controlled progress. The team proved value in one focused area, kept people in the loop, built governance into the work, and created a path to expand GenAI across more commercial banking journeys. That’s how AI becomes useful in the real world: not as a leap of faith, but as progress people can see and outcomes the business can trust.

Want to dive deeper into Lloyds Banking Group’s journey? Watch their PegaWorld session here.