◆ AI AT WORK · Workflows
Lloyds uses AI agents on debit-card fraud calls
The News
On October 26, 2026, Lloyds said an agentic AI platform that helps its fraud specialists investigate suspected debit-card fraud had supported them across more than 10,600 live customer cases in 90 days, and that its overall decision accuracy improved to over 90% in September. Trained specialists review the evidence and suggested actions and remain accountable for the customer decision.
Chandni Melwani
Founder & Editor
Oct 29, 2026 · 1 MIN READ
Know More
- — Systems per call: roughly halved; the process previously required information across 12 to 15 systems.
- — Fraud flags: in cases where it identified suspected fraud, more than 90% were subsequently confirmed as fraud.
- — Overrides: colleague overrides of the platform's recommended decisions have fallen by 30%.
When a customer contacts Lloyds about suspected debit-card fraud, a fraud specialist can work with a group of AI agents, each designed to support a different part of the investigation. A central system coordinates them and guides the specialist through up to nine stages of a case, according to Lloyds’ October 26 release.
Lloyds improved the platform while it was in use. The bank tested its outputs against completed case outcomes and used fraud specialists’ feedback to refine how it identifies relevant information, assesses fraud indicators and presents suggested actions. Over 90 days, seven updates delivered 50 improvements, which Lloyds says contributed to more consistent recommendations and fewer false alarms and missed fraud cases.
The figures are Lloyds’ own, and the release cites no outside review. “We will continue to track how accurately it identifies fraud as we develop the platform further,” said Rohit Dhawan, Lloyds’ group head of AI and advanced analytics.
The Read
Lloyds says the platform gathers more of the information specialists can need from multiple systems, 'allowing the specialist to focus on the customer.' The platform will probably become the specialist's main screen, since Lloyds aims for two or three systems per call by the end of 2026 and, over time, one AI-enabled workspace.
Room for Disagreement
Lloyds' planned AI-enabled workspace comes 'over time', and the release doesn't say that workspace will be this platform. Roughly halving 12 to 15 systems still leaves several on a call, and two or three by the end of 2026 is an aim.
Related
Frequently Asked Questions
Is this the agentic AI system Lloyds Banking Group announced in June?
The October 26 release doesn't say. On June 8, Lloyds Banking Group said several of its teams, including fraud, technology, data and risk specialists, had built an AI fraud agent using Envoy, the group's secure AI platform. It said multiple AI agents would operate simultaneously during customer calls, carrying out tasks such as identity checks, transaction analysis and scam risk assessment in real time, and that colleagues remain fully accountable for outcomes, with the ability to override AI suggestions.
Is Lloyds rolling the platform out further?
The release's first paragraph says Lloyds 'expands agentic AI platform to strengthen customer fraud support', and a later paragraph says Lloyds 'is continuing to test and monitor the platform before expanding its use'. It also says components developed for the platform are already being reused across other fraud journeys and could support further reuse, including in areas such as quality assurance.
How much is the fraud platform worth to Lloyds?
The release gives no figure for the platform alone. It calls the platform one example of how Lloyds is introducing next-generation AI across the business, and says that, taken together, 'these AI programmes are expected to contribute more than £100 million in additional value during 2026.' In June, Lloyds Banking Group said it prevented more than £1 billion of fraud in 2025 and has invested £100 million in new fraud technology since 2023.
Sources
- Lloyds, 'Lloyds uses agentic AI platform to strengthen customer fraud support', press release, October 26, 2026. Source for the 10,600 cases, 90 days, overall decision accuracy, the confirmed-fraud measure, overrides, systems per call, the agents and nine stages, the 50 improvements in seven updates, accountability, component reuse, the £100 million for AI programmes taken together, and Rohit Dhawan's quotes.
- Lloyds Banking Group, 'Lloyds Banking Group deploys agentic AI to strengthen real-time fraud protection', press release, June 8, 2026. Source for the June AI fraud agent, Envoy, the £1 billion of fraud prevented in 2025 and the £100 million invested in fraud technology since 2023.
Chandni Melwani
Chandni Melwani is the founder and editor of New in AI, covering AI agents, M&A, and enterprise adoption. She holds a Master's in Management of Artificial Intelligence from Queen's University and brings a practitioner's perspective from her work in Data and AI leadership.
Follow ↗AI at Work Briefing
Get the five-beat briefing in your inbox.
The big story, frontier labs, AI at work, e-commerce and the capital crossing borders, each with the part that tells you why it matters, not just that it happened.
Free · No spam · Unsubscribe in one click
More from AI at Work
Adoption
BofA says AI makes its tech development 10% more efficient
Bank of America CEO Brian Moynihan told analysts on October 14, 2026 that the bank has about 130 AI projects implemented, each with a payback, and that its technology development work is 10% more efficient.
Chandni Melwani · US · Oct 23, 2026
Enterprise
Citi says Devin helps it bring vendor work in-house
Citi CEO Jane Fraser told analysts on October 13, 2026 that 14,000 of the bank's engineers use Devin, its main agentic coding tool, and that in-sourcing work had cut its tech vendor spend 30% year over year.
Chandni Melwani · US · Oct 22, 2026
Jobs
Anthropic reports juniors fell to 9% of its technical hires
Anthropic's October 16, 2026 report on its own workforce says junior employees' share of its technical departments slipped from 21% to 14% since early 2025. It says the proportion of junior technical hires fell from 27% to 9% as hours worked by AI agents rapidly expanded relative to employees'.
Chandni Melwani · US · Oct 21, 2026