Key takeaways

  • The FRC published its Generative and Agentic AI Guidance on 30 March 2026, which it describes as the first such guidance from any audit regulator globally.
  • The FRC is clear that accountability does not change: firms and Responsible Individuals remain accountable for audit quality whatever tools they use.
  • ICAEW's Professional Standards Department said on 6 May 2026 that quality assurance reviews suggest AI use in audit work is currently very limited among the firms it monitors.
  • ICAEW says every audit firm should have AI policies and procedures, communicate them to staff and provide training, as part of its ISQM (UK) 1 quality management system.
  • EY's agentic AI rollout, launched in April 2026, covers 130,000 assurance professionals and more than 160,000 audit engagements worldwide.
  • In ICAEW's 2026 mid-tier research, 86% of firms said their technology strategy explicitly includes AI adoption, but only 17% felt confident assessing its workforce impact.
  • A Grant Thornton review of FTSE 100 audit reports as at June 2026 found no auditor had reported client AI use as a key audit matter.

UK audit firms are using AI in 2026 mainly to analyse full populations of journals, summarise documents such as board minutes and contracts, and run agents inside audit platforms that draft and organise work for human review. The largest networks have rolled agents out across their audit practices, while ICAEW says AI use among the smaller firms it monitors is still very limited. The FRC’s March 2026 guidance makes one point plain: the auditor stays accountable, whatever the tool.

How are UK audit firms using AI in 2026?

Most use falls into four areas, each with a human reviewing the output.

  • Journal and data analysis. Tools test every journal entry rather than a sample and flag anomalies, trends and risks.
  • Document review. AI summarises board minutes and reads contracts, for example to support revenue recognition testing.
  • Research. Agents answer questions on accounting standards and the firm’s audit methodology.
  • Workflow and drafting. Agents assign tasks, raise review notes, request client information and draft workpapers.

The FRC’s own illustrative examples include summarising board minutes and using an AI tool to review contracts for revenue testing. Its June 2025 publication looked at using AI to enhance procedures over journals.

What has the FRC said about AI in audit?

The Financial Reporting Council published its Generative and Agentic AI Guidance on 30 March 2026. It calls this the first such guidance from any audit regulator in the world, and its second publication on AI in audit.

DateFRC publicationWhat it covers
26 June 2025AI in audit: illustrative example and documentation guidanceAI over journals; documenting AI tools
30 March 2026Generative and Agentic AI GuidanceRisks to audit quality, mitigations, worked examples

The guidance sets out how to gain appropriate confidence in AI outputs. The FRC says the mitigations needed are a matter of professional judgement and depend on the tool and its use. Firms should fit AI use into their quality management system under ISQM (UK) 1.

Mark Babington, the FRC’s Executive Director of Regulatory Standards, said it is people, meaning the firms and Responsible Individuals, who are accountable for audit quality. Both documents sit in the FRC’s AI in audit library.

What has ICAEW said to smaller audit firms?

ICAEW’s Professional Standards Department published Artificial intelligence in audit work: managing the risks on 6 May 2026. It says quality assurance reviews suggest AI use in audit work is currently very limited.

Its main worry is staff using AI tools from their personal lives on client work without authorisation. That risks data protection breaches and the release of confidential client information, as well as harm to audit quality. ICAEW links AI to these rules:

  • ISQM (UK) 1: controlled AI use belongs in the firm’s quality management system
  • Audit Regulation 3.17: staff must be competent, including with AI tools
  • Audit Regulations 3.18 and 3.19: firms must have suitable audit procedures, and principals and staff must follow them

How are the largest firms using AI agents?

The big networks are building agents into their audit platforms. ICAEW Insights reported on 22 September 2026 that EY’s rollout, launched in April 2026, covers 130,000 assurance professionals in more than 150 countries and more than 160,000 audit engagements.

EY’s platform processes up to 1.4 trillion lines of journal data a year, and it expects agents to support end-to-end audit activity by 2028. Each auditor acts as the required human in the loop. EY UK partner Richard Harrison told ICAEW: “You can’t delegate authority to an artificial intelligence.”

ACCA’s AB magazine reported in August 2026 that KPMG has agents in its Clara platform, Deloitte launched its Zora agentic platform in 2025, and PwC says calendar 2026 audits will use an AI-integrated process. These are the firms’ own statements, not independent findings. Read more in our guide to AI agents in accounting.

What about mid-tier firms?

Mid-tier firms plan to adopt AI but are cautious about its effects. ICAEW’s third Evolution of mid-tier accountancy firms research, based on leaders from 35 firms in February and March 2026, found:

  • 86% have a technology strategy that explicitly includes AI
  • 71% think AI will let them move up the value chain
  • 68% expect AI to reduce the need for some early-career roles
  • only 17% feel confident assessing AI’s impact on their workforce

See AI and junior accountants.

How does client use of AI affect the audit?

The audit approach does not change; AI output is treated like any other management information. Writing for ICAEW on 24 September 2026, Grant Thornton’s Jon Morris said auditors should focus on where AI produces figures that feed the financial statements.

Higher-risk areas include expected credit loss models, stock provisions, valuation inputs and revenue judgements. These fall under ISA (UK) 315 risk assessment and ISA (UK) 540 on estimates. In his FTSE 100 sample as at June 2026, no auditor had reported client AI use as a key audit matter.

What audit firms should do now

  1. Write an AI policy. Say which tools are approved and for which tasks; our AI policy template is a starting point.
  2. Ban unapproved tools on client data. See is ChatGPT safe for client data.
  3. Train staff so they meet Audit Regulation 3.17 on competence.
  4. Record AI use in the file, following the FRC’s June 2025 documentation guidance.
  5. Review for hidden AI use when you review audit work.
  6. Tell clients how you use AI and how you protect their data.

Frequently asked questions

Can auditors use AI on UK audits?

Yes. The FRC supports the use of AI tools, but firms and Responsible Individuals remain fully accountable for audit quality and must judge how far to rely on any output.

What is the FRC’s guidance on generative AI in audit?

The Generative and Agentic AI Guidance, published on 30 March 2026, covers risks to audit quality, ways to gain confidence in AI outputs and worked examples such as summarising board minutes.

Do small audit firms need an AI policy?

ICAEW expects all audit firms to have AI policies and procedures, communicate them and train staff. This applies even if the firm uses AI very little.

Will AI replace auditors?

The regulators say no human accountability can be delegated to AI. ICAEW’s mid-tier research found 68% of firms expect AI to reduce the need for some early-career roles, not to remove auditors.

Sources

Rates, thresholds and rules in this guide were checked against HMRC and GOV.UK guidance on 6 November 2026. Tax rules change, so check the current position on GOV.UK or with a qualified adviser before acting. This guide is general information, not personal advice.