Lloyds Banking Group
Lloyds Banking Group's Athena GenAI search cuts staff lookup time by 66%
Lloyds Banking Group built Athena, a generative AI search tool that helps customer-facing colleagues find answers in the bank's internal knowledge articles. The bank says average search time fell by 66% and about 20,000 colleagues now use it.
Key figures
Copied exactly from the sources. Copy any figure with its citation.
66%
Reduction in average search time, as reported by Lloyds1
Around 20,000
Colleagues using Athena (February 2026 report)1
2.1 million
Searches by employees since the start of 2025 (July 2025 report)2
What it took
Cost, team size and timeline were not disclosed in the sources.
The problem
Customer-facing staff at a large bank must answer questions on every product and process. The answers sit in thousands of internal articles. Searching them takes time, and the customer waits on the line while the colleague looks.
What they built
Lloyds built Athena, a generative AI tool (software that writes answers in plain language) that searches the bank's internal knowledge hub. Finextra describes it as the bank's first major GenAI product. The hub holds 13,000 internal articles, and Athena covers support queries on the topics a customer might raise.
How they did it
Lloyds rolled Athena out to customer support colleagues, including telephone banking teams. Finextra reports the bank planned to add more users in relevant support roles during 2025. The sources do not describe the build team, cost, model or timeline.
Results
Lloyds says Athena cut average search time by 66%, from 59 seconds to 20 seconds. By February 2026 the bank said about 20,000 colleagues used it. It reported around £50 million of value from its generative AI work across the group in 2025, but that figure covers more than 50 solutions and is not specific to Athena.
All reported figures
| Measure | Figure | Citation |
|---|---|---|
| Reduction in average search time, as reported by Lloyds | 66%1 | |
| Colleagues using Athena (February 2026 report) | Around 20,0001 | |
| Average search time before and after (July 2025 report) | 59 seconds to 20 seconds2 | |
| Searches by employees since the start of 2025 (July 2025 report) | 2.1 million2 | |
| Time saved searching in telephone banking teams alone, as projected by Lloyds | 4000 hours2 |
Lessons for your project
Start with an internal use case. Lloyds' first major GenAI product served its own staff, which keeps risk lower than customer-facing answers.
Measure one simple thing before launch. Search time per query gave Lloyds a clear before-and-after number it could repeat in results.
Point the tool at content you already maintain. Athena searches an existing hub of internal articles, so answer quality depends on keeping those articles current.
Expand by role. Lloyds planned to add more customer support roles over time rather than launching to everyone at once.
Caveats
- All figures are Lloyds' own statements, relayed by trade press (Banking Exchange on the bank's results, Finextra on its announcement). No independent measurement is reported.
- The two sources give different user counts (about 20,000 in February 2026, 21,000 in July 2025) because they cover different dates and wording.
- The model and technology vendor are not named in the sources.
- No go-live date is stated. The deployed date shown is the date of the Finextra announcement article, although that article says employees had used Athena since the start of 2025, so it went live earlier.
- The 4000 hours saved is a projection stated by Lloyds for telephone banking teams, not a measured result.
- The Finextra page could only be read through an archived copy.
Sources
- Lloyds Says GenAI Delivered £50m of Value in 2025
Banking Exchange · 2026-02-04 · News · m.bankingexchange.com
- Lloyds trumpets GenAI knowledge hub's impact
Finextra · 2025-07-15 · News · finextra.com
Summarized in our own words from the sources above. Spotted an error? See corrections.