Big Four AI Consulting Alternative UK: The Auditor Test

A UK firm comparing a Big Four AI practice with an alternative usually starts with the wrong question. It asks who is bigger, who has the better platform, or who has the better brand. The question that settles more shortlists than any of those is simpler: which firm signs your annual accounts? In 2024 the Big Four audited 87% of FTSE 350 audit engagements, according to the FRC's Audit market and competition developments 2025 update. For most large UK companies, one of the four is the statutory auditor, and the auditor is the one Big Four firm whose AI work is restricted by audit ethics rules.

This guide is written for the Chief Data and AI Officer, or the transformation lead, who has to put an AI programme in front of a CFO and an audit committee. It explains what the FRC's Revised Ethical Standard 2024 says about an auditor building technology for its own audit client, how that differs for a listed or regulated public interest entity and a private mid-market company, and how to turn it into a buyer's grid. It deliberately cedes three neighbouring questions. Who actually delivers the work, and how to check, is covered in our post on Faculty after the Accenture deal. The general ranking of UK AI consultancies, and what consulting costs, sits in our enterprise AI consulting guide. Keeping the builder and the assurer of an AI governance system independent is covered in our AI governance consulting comparison.

This guide is shipped by the team behind DATS, the five-stage AI consulting system from DILR.AI, delivered by senior practitioners who ship code, not decks. Or see our AI operating model service, which covers governance, RACI and lifecycle, audit-ready by design.

What does a Big Four AI consulting alternative actually mean in the UK?

A Big Four AI consulting alternative is any firm other than Deloitte, EY, KPMG and PwC that can design and deliver an enterprise AI programme. The field includes Accenture, which is not one of the Big Four, mid-size technology consultancies, platform companies and senior-led boutiques. The useful comparison is not size against size, but which firms are free to build what a buyer needs.

The definition matters because the regulator uses it precisely. The FRC's 2025 audit market update defines the Big Four as Deloitte, EY, KPMG and PwC, and measures everyone else as non-Big Four. Accenture is a large consultancy rather than an audit firm, so the audit ethics rules discussed below do not reach it in the same way. That is one reason our enterprise AI consulting guide treats Accenture and the Big Four as a group for scale while this post separates them.

The alternative field is wide. It runs from Accenture through technology delivery firms such as Kainos, Made Tech and Slalom, to applied AI specialists such as Distyl, platform companies such as Palantir, and senior-led consultancies such as DATS. Faculty, long one of the best-known UK AI consultancies, now sits inside Accenture; what that change means for a client is the subject of our Faculty alternative guide. Each of these firms sells something different, so a buyer should first decide what the programme needs, using a structured AI placement method, and only then decide whom to ask.

Nothing in this guide questions the AI capability of the large firms. EY, for example, announced its EY.ai platform in September 2023 following a US$1.4b investment. Investment is evidence of budget and capability. It is not evidence that a particular Big Four firm is free to build a particular system for a particular client, and that freedom is where the comparison starts.

Why does your statutory auditor matter when you hire for AI?

Your statutory auditor matters because UK audit ethics rules restrict what an audit firm may build or run for a company it audits. The FRC's Revised Ethical Standard 2024 binds the audit firm, not the client, and it applies to all audit firms and all audits. If one of the Big Four signs your accounts, its AI practice faces limits that the other three, and every non-audit firm, do not.

The concentration in the audit market turns this from an edge case into a default. The FRC's 2025 update reports that the Big Four audited 87% of FTSE 350 engagements in both 2023 and 2024 and earned 98% of FTSE 350 audit fees in 2024. On a separate count, the FRC's Key Facts and Trends announcement of 15 October 2025 counted 48 FTSE 350 audits undertaken by non-Big Four firms, up from 41 in 2023, while the market share measure in the 2025 update held at 13%. If you are unsure how this applies to your own shortlist, talk to our team before the first proposal lands.

The practical effect is that, for most large UK companies, the shortlist of Big Four AI practices is really a shortlist of three for anything that touches the finance estate. For a company that is not a public interest entity, the fourth firm may still be able to help with work far from the accounts, under safeguards; for a public interest entity, the permitted list closes most of that route. Either way, the buyer should know which firm audits the company before the first scoping call. A CFO can answer the question quickly, and asking it first avoids proposal work on a firm that later declines or narrows the scope. If you are mapping a whole programme rather than one use case, the DATS placement diagnostic produces a ranked roadmap of where AI belongs and where it does not, which is the right place to record this constraint.

There is a second reason the auditor matters. Many AI programmes reach the finance function sooner or later: invoice processing, reconciliations, forecasting, revenue assurance, expenses and credit control are obvious candidates, and each of them sits close to the accounting system. The six enterprise AI solutions that DATS names, from post-call follow-up automation to developer productivity and the review queue, show how quickly operational AI ends up feeding a ledger somewhere.

What can your auditor's firm not build for you under the FRC Ethical Standard?

Under the FRC Revised Ethical Standard 2024, an audit firm shall not design, provide or implement IT systems for an audited entity where those systems would be important to any significant part of the accounting or financial management system, or to the production of the audited financial statements, and the audit team would place significant reliance on them. Nor may it do so where it would take part of the role of management.

The standard frames the risk in plain terms. Paragraph 5.49 of the FRC's Revised Ethical Standard 2024 states: "Design, provision and implementation of information technology (including financial information technology) systems by firms for an entity relevant to an engagement creates threats to the integrity, objectivity and independence of the firm and covered persons." The same paragraph names the two principal threats as the self-review threat and the management threat. An auditor that built your AI forecasting model would later be auditing figures its own model produced, which is the self-review threat in its simplest form.

Three practical consequences follow for an AI programme:

1. Finance-adjacent builds are the hard line. An AI system that posts journals, drives reconciliations, produces forecasts used in the accounts or feeds the financial statements is the kind of system paragraph 5.52 prohibits the auditor from designing or implementing, where the audit team would rely on it.
2. Running the system raises threats too. Paragraph 5.53 lists services that create threats, including hosting the audited entity's data, acting as the only access to its information systems, back-up and disaster recovery, and operating, maintaining or monitoring its IT systems. A managed AI service run by your auditor raises the same questions as the build, which is one reason buyers look at self-hosted AI options they control.
3. Taking decisions is caught everywhere. Where the IT engagement would mean the firm undertakes part of the role of management, paragraph 5.52 prohibits it whatever the system does, subject only to the conditional relief for small-entity audits in paragraph 6.11. An AI programme in which the consultancy chooses use cases, owns the model and approves its outputs edges towards that line.

For an audited company that is neither a public interest entity nor an other entity of public interest, work far from the accounts is treated differently. Paragraph 5.50 says IT systems that are not important to any significant part of the accounting system or the financial statements, and that the audit team does not rely on significantly, may be undertaken where appropriate safeguards can be applied. Paragraph 5.51 gives examples of those safeguards: staffing the IT project with partners and staff who have no involvement in the audit, and a review by an uninvolved partner with relevant expertise. A customer-service assistant or a knowledge-retrieval tool may therefore be possible for an audit client, but only through the firm's own independence process, and the decision sits with the audit firm, not the buyer.

The same diagnostic logic underpins our AI operating model consulting, which covers governance, RACI and lifecycle, audit-ready by design, so that your own people stay accountable for the decisions an AI programme raises.

Does the auditor restriction apply to a private mid-market company?

Yes, in part. The FRC Ethical Standard applies to all audit firms and all audits, so the prohibition on building finance-critical IT systems that the audit relies on reaches a private mid-market company audited by a Big Four firm. The permitted-list regime and audit committee approval apply to public interest entities and other entities of public interest, and the 70% fee cap to public interest entities.

The difference between the two regimes is large, so scope it carefully before reading anything into a proposal, as part of a wider AI readiness check.

For a public interest entity, the rule is framed as a permitted list. Paragraph 5.40 says an audit firm carrying out statutory audits of public interest entities, and any member of its network, shall not provide the audited entity, its UK parent or its worldwide controlled undertakings with services other than those on a permitted list, subject to the approval of the audit committee, and paragraph 5.43 applies the same requirements to an other entity of public interest. The permitted list centres on reporting required by law or regulation, interim reviews and assurance work closely linked to the audit. Designing an AI system is not on it; assurance of AI governance is a separate question, covered in our AI governance consulting comparison.

For a private company, the picture depends on the system. Both kinds of buyer are common: a UK bank or insurer is a public interest entity, while a privately owned manufacturer or a mid-size professional services firm usually is not. A bank or insurer also answers to the FCA and the PRA for how it governs AI, and those expectations sit alongside the audit rules rather than inside them. Our mid-market AI readiness guide covers how a firm of that size can scope its first use cases before any of this becomes a procurement question.

Section 6 of the standard, which sets out provisions available for audits of small entities, offers some alternative provisions and exemptions. It does not apply to public interest entities, and whether a dispensation covers a given service is a question for the audit firm's ethics partner. A buyer should treat any such conversation as the auditor's to resolve, in writing, before work starts.

How does an auditor conflict change the timing of an AI roadmap?

An auditor conflict changes the timing of an AI roadmap in three ways for a public interest entity. Every non-audit service needs audit committee approval before it starts, total non-audit fees are capped against the audit fee, and some prohibited services, including designing financial IT systems, also block a firm in the financial year before it begins an audit.

The approval step comes first. Paragraph 5.30 of the Revised Ethical Standard 2024 requires an audit firm carrying statutory audits of public interest entities or other listed entities to seek approval from those charged with governance before providing non-audit services, and to say whether the proposed service complies with the standard. A programme that needs an audit committee slot before a statement of work can be signed runs on the committee's calendar, not the programme's.

The fee cap follows. Paragraph 4.13 limits the audit firm's non-audit fees from a public interest entity audit client, outside services required by law or regulation, to no more than 70% of the average audit fees of the last three consecutive financial years. An AI build is not on the permitted list in the first place, so the cap bites on the work an auditor may still do around an AI programme, such as permitted assurance over controls, which shares the same headroom. Footnote 54 of the standard notes that an Other Entity of Public Interest is not subject to the 70% cap, which is one more reason to confirm the company's category with the CFO rather than assume it.

The third effect is the one buyers most often miss. Appendix B of the standard lists prohibited non-audit services for public interest entities, and item (e) covers designing and implementing internal control or risk management procedures related to financial information, or designing and implementing financial information technology systems. The prohibition runs from the beginning of the period audited to the audit report, and, for items (e) and (h), through the financial year immediately before that period. For a public interest entity, a firm that builds a finance AI system for you this year may therefore be ineligible to become your auditor next year. If an audit tender is on the horizon, hiring a Big Four firm for finance-adjacent AI can quietly remove it from the tender, so raise the tender timetable at the scoping stage.

One further point concerns visibility rather than timing. Paragraph 5.38 notes that UK legislation requires large companies to disclose fees receivable by their auditors and their auditors' associates, by category. Material non-audit work by your auditor's firm therefore appears in the accounts, and a buyer should be ready to explain it. Our other AI strategy guides cover how boards review AI programmes more broadly.

What should a buyer's grid compare between a Big Four practice and an alternative?

A buyer's grid for a Big Four AI practice and an alternative should compare independence before capability: whether the firm audits you, whether the AI will touch the finance estate, who will run the system afterwards and who takes the decisions. Capability, reach and delivery model come next. The grid below turns the FRC rules, and the commercial checks that follow them, into questions a procurement team can ask.

Two cautions apply to the independence rows. First, the rules bind the audit firm, so a Big Four firm that is your auditor is responsible for declining or reshaping work it cannot accept, and it should raise the issue itself. Second, the grid does not tell you which firm is better. It tells you which firms are free to do which work, so that the comparison of capability happens on a fair field.

Ownership is the row most often left to the end. A system that a consultancy hosts and runs is harder to move later, whoever built it. Our guide to self-hosted AI for regulated institutions covers how to keep models and data inside a perimeter you control, and the AI execution office is how DATS runs embedded delivery with production placements the client owns.

Delivery model is the other row that separates firms more than brand does. Some buyers want a project team that hands over and leaves; others want a small embedded team that works alongside their own people for a year or more. Our comparison of an embedded AI delivery team and project consulting covers when each fits, and the DATS five-stage method runs from Discover and Diagnose through to Scale and Run.

What is the best Big Four AI consulting alternative in the UK for 2026?

The best Big Four AI consulting alternative in the UK for 2026 depends on two facts: who audits you, and how close the AI work sits to your accounts. For a finance-adjacent programme at a company audited by a Big Four firm, the strongest shortlist is drawn from capable firms that are not your auditor. For multi-country rollouts with no audit conflict, a Big Four practice can be the right call.

That concession is real, not a courtesy. A Big Four firm that is not your auditor brings a global bench, sector teams and the ability to staff work in many jurisdictions at once, and for a multinational standardising AI across dozens of entities, that scale can outweigh everything else in the grid. Accenture competes on the same axis without the audit question at all, and our Faculty alternative guide covers what its ownership change means for clients. If the work is far from the accounts and your company is not a public interest entity or an other entity of public interest, even your own auditor's firm may be able to help under safeguards, provided it confirms in writing how it has applied the standard.

An alternative is worth a hard look in three situations. The first is the finance-adjacent programme at an audit client, where the auditor is constrained and the other three Big Four firms may be bidding for your next audit. The second is the mid-market buyer that needs senior attention and a short path from scope to a shipped system, rather than a programme sized for a FTSE 100 group. The third is the buyer who wants production placements it owns rather than a system an adviser keeps running.

For that last group, a senior-led system such as DATS fits well. DATS works in three productised engagements: a four to six week placement diagnostic that produces a ranked roadmap of where AI belongs and where it does not, operating model design over six to ten weeks, and a twelve-month-plus execution office for embedded delivery. Read more about who we are and read about the DATS approach to AI placement. As with any adviser, a buyer should test those claims against the grid above rather than take them on trust.

Frequently asked questions

Is Accenture one of the Big Four?

No, Accenture is not one of the Big Four. The FRC defines the Big Four as Deloitte, EY, KPMG and PwC, the four firms that dominate UK statutory audit. Accenture is a large consultancy rather than one of the audit firms the FRC counts in the Big Four, so the audit ethics limits discussed here do not bind it as your statutory auditor. It is a frequent alternative to a Big Four AI practice.

How Accenture and the Big Four compare on scale is covered in our enterprise AI consulting guide.

Can our auditor still advise us on AI at all?

Your auditor can still provide some AI-related work, but the FRC Ethical Standard sets the limits. For a public interest entity, or an other entity of public interest, the audit firm may provide only services on the permitted list, with audit committee approval. For other audited companies, the firm assesses threats and applies safeguards, and work far from the accounts may be possible. Building finance-critical systems that the audit relies on is prohibited for every audit client.

Where the question is assurance of an AI governance system rather than a build, our AI governance consulting comparison explains how to keep the builder and the assurer apart.

When did the current FRC Ethical Standard take effect?

The FRC published its Revised Ethical Standard 2024 on 15 January 2024, and it took effect from 15 December 2024. For public interest entities, it frames the auditor's non-audit work as a list of permitted services that need audit committee approval. Any AI engagement starting in 2026 with a Big Four audit client falls under the 2024 standard.

The FRC's announcement of the update summarises what changed, and our workforce evidence review covers the people side of the same programmes.

Want to see this in practice? Read how DATS senior-led AI consulting works, explore our enterprise AI solutions, read about our approach to AI placement, or browse more AI strategy guides.

Written by the Dilr.ai engineering team, practitioners who ship enterprise AI in production. Follow us on LinkedIn for shipping notes, or subscribe via the RSS feed.