Lloyds Banking Group AI Strategy: Agentic AI at Scale, with a Value Target

Lloyds Banking Group is the highest-ranked UK bank in the Evident AI Index, in 15th place. It has named five areas where agentic AI will change the bank. It is also one of the few banks to put a money value on AI: about £50 million in 2025, with more than £100 million targeted for 2026.

The ambition is agentic AI in practice. In a January 2026 post, Rohit Dhawan, Group Director of AI and Advanced Analytics, wrote that 2026 is the year agentic AI moves from promise to practice.

The focus is five named areas. The same post lists:

• customer interactions
• back-office operations, such as fraud investigation, complaints and credit support
• frontline and relationship manager support
• colleague assistance, including agent-building tools and a colleague “super-agent”
• engineering and developer enablement

AI now sits inside the bank’s cost plan. In its July 2026 half-year results and strategy update, Lloyds targeted about £2 billion of savings by 2030. Technology including AI is behind them, as reported by Reuters via Resultsense.

Lloyds has not said publicly what it will not prioritize.

Agents now run on a shared platform. Lloyds unveiled Envoy, a platform for building, sharing and running AI agents, in April 2026, according to its AI page. Agentic fraud detection that analyses payments in real time followed in May 2026.

Generative AI is already in production. The bank says it deployed more than 50 generative AI solutions in 2025 (press release, January 2026).

The technology is built with partners. Lloyds uses Microsoft 365 Copilot and GitHub Copilot. It also runs an AI Centre of Excellence with more than 200 specialists.

The workforce stance favors new roles. Lloyds says agentic AI will automate routine tasks and create new roles in oversight and strategy.

In June 2026, it announced more than 1,000 AI roles for 2026. That is nearly 300 new hires and more than 700 existing colleagues moving into AI work, alongside new AI engineering apprenticeships.

In July 2026, CEO Charlie Nunn said the bank does not set targets for staff numbers.

Enablement is the bank’s strongest area. Every colleague takes mandatory responsible AI training through a group-wide AI Academy, which also offers role-specific learning.

A volunteer network of about 1,000 “flight instructors” coaches colleagues on Copilot, a working example of AI champion programs at scale. More than 10,000 people had been trained by 2025, according to a Microsoft case study.

Ownership changed hands in 2026. Sameer Gupta joined from DBS as Chief Data and AI Officer in June 2026, as reported by Banking Dive. Rohit Dhawan leads the AI program as Group Director of AI and Advanced Analytics. No AI-specific investment figure has been disclosed.

Governance is described in practice. A Responsible AI team applies a framework of rigorous testing, human oversight and scenario planning, with real-time monitoring and escalation for agents.

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

• Fluency floor: Level 1 (Chatting). Every colleague must complete responsible AI training. That sets a safe-use baseline but not a work-quality standard.
• Highest level evidenced: Level 3 (Integrating), company-claimed. About 5,000 engineers using GitHub Copilot report a 50% improvement in converting legacy code, which is a named, recurring workflow with a measured result. Agents are live, but none of their results have been published, so Level 4 is not evidenced.
• Portfolio design: Multi-level. Role-specific learning, a champion network and specialist AI roles run side by side. No role-specific targets have been disclosed.

It puts a money value on AI. Lloyds reports about £50 million of value from generative AI in 2025 and targets more than £100 million in 2026. Both figures are company-claimed.

AI is tied to revenue as well as cost. At its November 2025 Digital and AI Investor Seminar, the bank said digital and AI initiatives would deliver more than 70% of its 2026 strategic revenue target.

Adoption is high where tools are licensed. In October 2025, 93% of colleagues with Microsoft 365 Copilot licences were actively using it, out of nearly 30,000 licences. A colleague survey put the time saved at 46 minutes a day.

Two results are still missing: performance against the 2026 target, and results from the agents.

Change log

• March 2025: Lloyds commits to 40,000 Microsoft 365 Copilot licences.
• October 2025: Nearly 30,000 licences are deployed, with 93% actively used.
• November 2025: An agentic financial assistant is announced, and the Digital and AI Investor Seminar is held.
• January 2026: £50 million of value for 2025 is reported, more than £100 million is targeted for 2026, and the five focus areas are published.
• April 2026: The Envoy agent platform launches.
• May 2026: Agentic fraud detection is deployed.
• June 2026: More than 1,000 AI roles are announced, and a new Chief Data and AI Officer joins.
• July 2026: The strategy to 2030 targets about £2 billion of savings, with AI among the drivers.
• October 2026: Evident ranks Lloyds 15th, the highest of any UK bank.

• Disclosure: 4 of 4. Lloyds publicly answers all four questions, including a value figure, a one-year target and named focus areas.
• Measurement is its strength. It reports AI value in money and has set a target for the next year.
• The people model is unusually detailed. Mandatory training, a 1,000-strong champion network and more than 1,000 AI roles give it a clear enablement structure.
• The say-do gap is in agents. Envoy and the fraud-detection agents are live, but their results have not been published, and Evident ranks Lloyds 27th on innovation.
• Watch the 2026 number. The bank’s own £100 million target is the clearest test of whether the strategy is working.

What the numbers show: Revenue per employee has barely moved: up 4% in four years, with headcount slightly above 2022. Lloyds is hiring for AI roles rather than cutting, and no productivity effect is visible at group level yet.

Revenue, headcount and revenue per employee, FY2022 to FY2025

• FY2022: £17.46bn revenue, 63,372 employees, £276k per employee
• FY2023: £17.93bn revenue, 66,197 employees, £271k per employee
• FY2024: £17.12bn revenue, 66,229 employees, £258k per employee
• FY2025: £18.30bn revenue, 64,038 employees, £286k per employee

Change, FY2022 to FY2025: revenue up 5%, headcount up 1%, revenue per employee up 4%.

Notes: Revenue is Lloyds’ underlying net income, its headline income figure. Headcount is the average number of employees during the year. 2022 is restated for the IFRS 17 insurance accounting standard. Trend figures stay in £ so exchange-rate moves do not distort them. At the 2025 average exchange rate, latest revenue per employee is about $377k.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: Lloyds results announcements and 20-F filings.

What is Lloyds Banking Group’s AI strategy?

Lloyds is moving agentic AI from pilots to enterprise scale across five areas: customer interactions, back-office operations, frontline support, colleague assistance and engineering. It pairs this with a large colleague training program and a value target for 2026.

‍

How much value does Lloyds get from AI?

Lloyds reports about £50 million of value from generative AI in 2025 and targets more than £100 million in 2026. Both figures come from the bank. It also links AI to about £2 billion of planned savings by 2030.

‍

How many Lloyds colleagues use AI?

In October 2025, 93% of colleagues with Microsoft 365 Copilot licences were actively using it, out of nearly 30,000 licences. All colleagues complete mandatory responsible AI training.

‍

Is Lloyds cutting jobs because of AI?

Lloyds says AI will automate routine tasks and create new roles, and it announced more than 1,000 AI roles for 2026. Its CEO has said the bank does not set targets for staff numbers.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings materials, filings and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix. Company figures are marked as company-claimed unless they have been independently verified.