AI-search summary

HM Revenue & Customs is expanding its use of artificial intelligence and data analytics to enforce existing tax rules. A £175m, ten-year Quantexa contract, 5,500 extra compliance officers and a stated focus on wealthy and offshore taxpayers accompany a tax gap that HMRC now estimates at 6.4% for 2024-25.

Key takeaways

  • HMRC agreed a £175m, ten-year data and AI contract with British firm Quantexa, announced on 14 May 2026, to modernise its data systems and scale governed AI.
  • HMRC estimates the 2024-25 tax gap at 6.4%, or about £59.2bn, up from a revised 6.0% (£52.8bn) in 2023-24; small businesses account for the largest share.
  • The department is recruiting 5,500 additional compliance officers by 2029-30 and expects government measures to raise about £10bn a year in extra tax by then.
  • Wealthy and offshore compliance is a stated priority, with a Further Closing the Wealthy Tax Gap plan due in Autumn 2026; this concerns enforcement of existing law, not proven wrongdoing.
  • Making Tax Digital for Income Tax began in April 2026 for sole traders and landlords earning over £50,000, widening HMRC's real-time data and its analytical reach.

Britain's tax authority is leaning ever more heavily on artificial intelligence and data analysis to enforce the rules it already has. HM Revenue & Customs has agreed a £175m, ten-year contract with the British technology firm Quantexa, announced on 14 May 2026, to modernise its data infrastructure and expand what it describes as sovereign, governed AI at national scale. The move sits within a wider transformation programme that includes thousands of additional compliance officers and a declared focus on wealthy and offshore taxpayers. It arrives at a moment when HMRC's own estimate of the tax gap, the difference between tax owed and tax collected, has risen rather than fallen. None of this implies wrongdoing by any individual; it reflects an effort to collect tax that is already legally due.

What has been announced

The clearest single development is the Quantexa contract. According to the company's own statement, the ten-year, £175m arrangement is intended to modernise HMRC's data infrastructure, join up fragmented information and enable advanced analytics to support decision-making and strengthen controls. HMRC has framed the deal as part of building a more capable, technology-led compliance operation rather than as a new power to tax.

Around that contract sits a broader set of commitments. HMRC's Transformation Roadmap progress update, published in 2026, describes testing AI-powered systems that give caseworkers rapid and consistent guidance, piloting automated call and case summarisation, and developing tools to help detect forged documents during investigations. The department has also carried out one of the largest rollouts of Microsoft Copilot in UK government, with more than 28,000 licences distributed by March 2026 and plans to reach around 50,000 users.

Crucially, HMRC has been at pains to stress that technology supports rather than replaces human judgement. The department has said that final decisions on individual cases are made by experienced, trained caseworkers. That distinction matters, because much of the public anxiety about automated enforcement concerns whether a computer, rather than a person, is deciding who is investigated.

The tax gap that frames the effort

The backdrop to all this is a widening estimated tax gap. On 23 June 2026, HMRC estimated the 2024-25 tax gap at 6.4% of total theoretical liabilities, equivalent to roughly £59.2bn. At the same time it revised the 2023-24 figure upwards, to 6.0%, or about £52.8bn, from the 5.3% (£46.8bn) previously reported. These are estimates, subject to revision, but the direction of travel has drawn political attention.

The composition of the gap is often misunderstood. HMRC's analysis attributes the largest single share to small businesses, at around 62% of the total, with Corporation Tax a significant component. Failure to take reasonable care, described as carelessness rather than deliberate evasion, is identified as the biggest behavioural driver. Evasion accounts for a smaller portion. In other words, much of the gap reflects error and complexity rather than concealment by the very wealthy.

That context is worth holding onto when interpreting talk of a crackdown. A rising estimated gap gives the government a fiscal incentive to invest in enforcement, but the numbers also caution against assuming that any one group is the sole cause. The wealthy are a stated focus, yet they are one part of a much broader compliance challenge.

How HMRC's data machine works

HMRC's analytical capability is not new. For well over a decade the department has operated data-matching and risk-analysis tools, most publicly associated with its Connect system, which draws together information from tax returns, banks, other government departments and a range of third parties to flag inconsistencies. The current wave of investment is best understood as an upgrade to that long-standing approach rather than a sudden departure from it.

The direction is towards richer, more connected data and faster analysis. HMRC is legislating to place new obligations on third-party data holders to improve reporting of financial account and card-sales information from 2027-28 onwards. It is also joining international agreements on the automatic exchange of information, including a cryptoasset reporting framework and arrangements covering overseas real estate, which extend its visibility into assets held abroad.

The Quantexa contract fits this pattern. Entity-resolution technology, the discipline of working out when different records refer to the same person or business, is central to turning scattered data into a coherent picture. Applied at scale, it can help identify undeclared income or hidden connections, but it can also, in principle, reduce mistakes by giving officers a fuller view before they act.

International cooperation is an increasingly important part of the toolkit. Automatic exchange of financial-account information between tax authorities has, over the past decade, sharply reduced the scope for income and assets to remain invisible across borders. The newer frameworks for cryptoassets and overseas property extend that reach into areas that were previously harder to see, giving HMRC more raw material to analyse. The department's investment in AI is, in part, a response to this abundance of data: the challenge has shifted from obtaining information to making sense of it quickly and accurately.

Why the wealthy and offshore are in focus

The government has made the wealthy and offshore compliance a stated priority. HMRC has reported recruiting external experts and redeploying experienced staff to work on offshore non-compliance by wealthy individuals, and it has committed to publishing a Further Closing the Wealthy Tax Gap plan in Autumn 2026. These are enforcement measures aimed at ensuring existing rules are followed, not new taxes.

There are practical reasons for the emphasis. Wealthy taxpayers tend to have more complex affairs, spanning multiple jurisdictions, trusts, partnerships and investment structures, which makes accurate assessment harder and mistakes, whether innocent or otherwise, more consequential in cash terms. Better data and analytics are intended to help HMRC understand these arrangements more fully rather than to presume fault.

It is important to frame this carefully. Holding assets offshore or arranging one's affairs efficiently is not, in itself, unlawful. The compliance effort is directed at cases where tax that is legally due has not been paid, and the vast majority of wealthy taxpayers meet their obligations. The stated aim is fairness and accuracy across the system, not the targeting of individuals for legitimate behaviour.

Making Tax Digital widens the net

Running alongside the enforcement drive is the long-planned Making Tax Digital programme. From April 2026, Making Tax Digital for Income Tax began applying to sole traders and landlords with qualifying income over £50,000, requiring digital record-keeping and quarterly updates to HMRC. The department has reported that hundreds of thousands of taxpayers signed up in the early phase.

The threshold is due to fall over time, reaching income above £30,000 from April 2027 and above £20,000 from April 2028, gradually bringing more taxpayers into the digital regime. For HMRC, more frequent digital reporting means more timely data, which in turn feeds the analytical systems used to spot anomalies. The programme is presented as a modernisation of record-keeping, but its compliance implications are clear.

Why it matters for businesses, individuals and the economy

For businesses, particularly smaller ones, the practical message is that record-keeping and accuracy matter more than ever. As HMRC's data becomes more connected and its analysis faster, discrepancies are more likely to be noticed. That is not a reason for alarm for compliant firms, but it does raise the premium on good systems, professional advice where affairs are complex, and prompt correction of genuine errors.

For individuals, the effect is mostly indirect but real. A more effective tax authority influences the fairness of the system and the amount of revenue available for public services. For the wider economy, improved compliance can, in principle, allow a given level of spending to be funded with less pressure to raise headline rates, although that is a political choice rather than an automatic outcome.

There is also a fiscal dimension that bears on the public finances. HMRC has said the government's measures are expected to deliver around £10bn in additional tax revenue each year by 2029-30. If achieved, that would be a meaningful contribution at a time of tight budgets, though such projections depend on delivery and on the economy performing broadly as assumed.

People, not just machines

The enforcement push is as much about staff as software. HMRC has committed to recruiting 5,500 additional compliance officers by 2029-30, and its progress update reported that more than 2,100 had already been hired, with further intakes planned. Compliance yield, the extra tax secured through this work, reached a record figure of around £48bn in 2024-25, with a higher target set for 2025-26.

The Autumn Budget 2025 added further measures. These included the creation of a small-business evasion and enforcement team drawing on hundreds of criminal investigators, additional investment in private-sector debt collection partnerships, enhanced whistleblower rewards for information leading to large recoveries, and tougher penalties in areas such as late Corporation Tax filing. Together they signal a broad tightening across the compliance landscape rather than a single measure aimed at one group.

Recruitment on this scale carries its own challenges. New compliance officers require training and experience before they become fully productive, and independent analysts have warned that the pace of hiring, alongside retention and skills, will influence whether the projected revenue is realised. The combination of more people and better tools is intended to be mutually reinforcing: technology helps officers focus their attention, while human judgement provides the checks that automated systems cannot. Whether that balance is struck well in practice is one of the central questions hanging over the strategy.

Key risks and uncertainties

Several risks attend this strategy. The first is accuracy and fairness: automated risk-flagging can generate false positives, and safeguards, transparency and effective appeal routes are essential if trust is to be maintained. HMRC's insistence that trained officers make final decisions is a safeguard, but the volume and complexity of automated analysis will keep questions of oversight and data protection prominent.

The second is delivery. Large technology contracts and ambitious recruitment targets do not automatically translate into results, and HMRC has a long history of complex modernisation programmes. Independent commentators have cautioned that staffing and training bottlenecks could slow the realisation of projected revenue. The £10bn-a-year figure and the compliance-yield targets are goals, not guarantees.

A third risk is proportion and perception. If enforcement is seen as bearing down hardest on small businesses making honest mistakes, or as unfairly singling out particular groups, public support could fray. The estimated tax gap data, which places much of the shortfall with small businesses and attributes a large share to carelessness rather than evasion, is a reminder that the story is more nuanced than a simple pursuit of the wealthy.

What to watch next

Several markers will show whether the strategy is working. The Further Closing the Wealthy Tax Gap plan, due in Autumn 2026, should set out the government's approach in more detail. Future tax-gap estimates will indicate whether the recent rise reverses. And HMRC's annual report and compliance-yield figures will reveal whether extra staff and technology are delivering the promised revenue.

It will also be worth watching how the Quantexa systems bed in, how Making Tax Digital's lower thresholds affect smaller taxpayers, and whether the safeguards around automated decision-making satisfy Parliament, professional bodies and the courts. The credibility of the enforcement drive will rest not on the technology itself, but on whether it is seen to be accurate, proportionate and fair.