, ,

HMRC’s roadmap brings AI into the tax service queue

HMRC has updated its transformation roadmap with digital service, contact centre, and AI plans for 2026 to 2027.

HMRC’s roadmap brings AI into the tax service queue
Summary
  • HMRC’s 2026 update sets out work on digital accounts, expenses, Child Benefit tracking, forms digitisation, and contact centre technology.
  • The department plans a CCaaS platform with embedded AI capability, alongside expanded digital self service.
  • The roadmap links public sector productivity to execution risk, accessibility, tax compliance, and trust in high volume services.

HM Revenue and Customs has updated its transformation roadmap, setting out how it plans to expand digital services, introduce new contact centre technology, and use AI across parts of tax administration during 2026 to 2027.

The update covers planned changes to the Personal Tax Account, a new digital expenses service, Child Benefit claim tracking, further forms digitisation, and work to make more services digital by default for users who already interact with HMRC online. The department says 78% of customer interactions are now digital, while the HMRC app added 2.8 million new users in 2025 to 2026.

HMRC also plans to begin work on a new cloud contact centre as a service platform with embedded AI capability. The system is intended to manage call queues, provide more accurate wait times, and support intelligent digital assistance across phone and online channels.

The department’s scale makes the roadmap an operational technology story as much as a government service update. Tax administration affects individuals, small companies, employers, advisers, and larger businesses that depend on predictable systems for compliance, refunds, disputes, and account management. When services are slow or hard to use, the cost is pushed into time, uncertainty, professional fees, and delayed decisions across the economy.

The roadmap shows HMRC trying to move more interactions into digital channels while keeping phone support for people who need it. That balance has been difficult before, especially when service performance has drawn criticism from taxpayers and professional bodies. Digital self service can reduce pressure on contact centres only when online services are complete enough to resolve the problems that cause calls in the first place.

Several of the planned changes are practical rather than dramatic. Better digital accounts, online evidence upload, claim tracking, pre-populated information, and more digitised forms may not sound ambitious beside large AI programmes, but high volume friction points are often where public sector productivity is won or lost. Removing a common avoidable call can have more operational value than launching a speculative pilot.

The AI element needs careful execution. Queue management, digital assistants, call summarisation, and guidance tools can improve service delivery when the underlying data and process design are strong. In tax, however, mistakes carry financial and legal consequences. Automated systems need clear escalation routes, audit trails, human accountability, and careful wording around what users can rely on.

Security and identity sit in the same programme. HMRC has completed test and learn activity on multi-factor authentication for tax advisers and intends to move towards broader operation during 2026 to 2027. Tax systems are attractive fraud targets, while advisers need secure access to client information without clumsy processes that slow legitimate work.

The update also sits within the government’s wider digital government agenda, which aims to make public services more joined up by 2030. HMRC will be one of the clearest tests of that ambition because it combines high transaction volumes, legacy systems, political scrutiny, complex rules, and users with very different levels of digital confidence.

Execution will decide whether the roadmap becomes service improvement or another layer of complexity. The strongest elements are tied to specific user journeys and operational bottlenecks. The weakest risk is familiar across public sector technology: making digital the default before the service is strong enough to carry the load.