How Lloyds is Using Agentic AI to Fight Fraud In Real Time

Lloyds Banking Group is scaling its use of AI to tackle one of financial services’ most persistent and costly challenges: fraud. 

With more than £1bn (US$1.3bn) of fraud prevented in 2025 and £100m (US$133.8m) invested in new fraud technology since 2023, the group is at the forefront of AI-driven financial crime prevention.

Scaling AI-led fraud protection

At the core of Lloyds’ strategy is a newly deployed agentic AI system designed to enhance real-time decision-making across its fraud operations. 

Built using the Group’s secure AI platform, Envoy, the system has been developed collaboratively by fraud, technology, data and risk teams.

Unlike traditional rule-based systems, agentic AI introduces multiple autonomous agents working simultaneously behind the scenes. 

These agents perform tasks such as identity verification, transaction monitoring and scam risk analysis in parallel, significantly accelerating response times.

Crucially, the system is embedded into tools already used by frontline colleagues, allowing them to act quickly while maintaining oversight. 

This approach ensures that technology augments, rather than replaces, human judgement.

Agentic AI with human accountability

Lloyds is emphasising a human-in-the-loop model, where colleagues retain full accountability for decisions. 

AI-generated insights can be overridden, ensuring regulatory alignment and customer protection remain paramount.

During customer interactions, particularly in high-risk scenarios, the agentic system enables faster, more informed interventions.

By surfacing relevant insights in real time, it allows staff to focus on supporting customers rather than navigating complex data sets.

Ron van Kemenade, Group COO at Lloyds Banking Group, says: “AI is now integrated across our organisation and is transforming how we operate and serve customers. 

“Our ongoing investment keeps Lloyds Banking Group ahead in responsible tech adoption within financial services.”

Real-time scam detection for customers

Alongside internal tooling, Lloyds is rolling out Scam Check, a customer-facing feature embedded within payment journeys across Lloyds, Halifax and Bank of Scotland.

The tool introduces a proactive layer of protection at the point of transaction. 

When customers attempt to pay a new recipient, Scam Check assesses the likelihood of fraud using machine learning and image analysis. 

If a risk is detected, users are prompted with tailored questions and asked to upload screenshots of the item they intend to purchase.

This capability is particularly relevant for marketplace and social media transactions, where authorised push payment scams are prevalent. 

By intervening before a payment is completed, Lloyds aims to shift fraud prevention upstream.

Making fraud protection visible

In a move to build customer trust, Lloyds has also introduced the Dark Horse logo within its banking apps and payment journeys. 

The symbol signals that fraud detection systems are actively monitoring transactions in the background.

This visibility strategy reflects a broader industry trend towards transparency in security measures, helping reassure customers without adding friction to the user experience.

“Fraud is increasingly fast-moving and complex, so our focus is on responsibly applying AI in a way that genuinely helps customers, without delaying prevention,” Tom Martin, Business Platform Lead for Economic Crime Prevention for Lloyds Banking Group, says.

“By combining agentic AI with human oversight, we can create an invisible forcefield around our customers, monitoring risk in real time and stepping in only when it matters.”

Responsible AI at scale

Underpinning these innovations is Envoy, Lloyds’ enterprise AI platform designed to enable secure, compliant and scalable deployment of AI solutions. 

The platform supports a wide range of use cases, from fraud detection to customer service enhancements, while ensuring alignment with regulatory and ethical standards.

Envoy’s architecture allows seamless integration with existing systems, enabling rapid development and deployment without compromising security. 

It also reinforces governance frameworks, ensuring that AI adoption remains controlled and auditable.

As financial crime grows more sophisticated, Lloyds’ investment in agentic AI reflects a broader shift across the industry: from reactive fraud detection to proactive, real-time prevention powered by intelligent systems and human oversight.