Want to know more?
Are you interested in this project? Or do you have one just like it? Get in touch. We’d love to tell you more about it.
How Equal Experts helped HMRC safely deepen its use of AI with built-in cost control
How a custom-built platform addressed AI barriers to help civil servants engage confidently with emerging technologies while optimising the value of AI spending.
HMRC wanted to help data scientists, developers and analysts to benefit from emerging AI technologies, while recognising the difficulty of using public large language models (LLMs) and protecting sensitive taxpayer data. The organisation partnered with Equal Experts to deliver a platform that could meet these needs and prove its value in pounds and hours saved.
Equal Experts supported HMRC to vastly extend an existing platform called the generative AI landing zone, or GAILZ. This included targeted product management and technical leadership to raise awareness of the platform across HMRC, streamline user onboarding, and embed granular financial visibility and control from the outset.
Today, GAILZ is housed entirely within HMRC’s private multi-cloud environment, ensuring total data sovereignty and paving the way for advanced AI adoption across the organisation.
Equal Experts supported HMRC to create an interactive AI training portal (EZAI) where staff can practice AI techniques. This includes a suite of ready-made utility apps with AI capabilities, known as RAIN apps. This project directly supports cross-government efficiency targets and talks are underway to extend the model across Whitehall as the blueprint for safe, high-impact public sector AI.
Outcomes
Access to secure AI tools
HMRC colleagues in data science, development and analyst roles have access to secure AI tools that support improved efficiency in key tasks.
Real-time visibility
of AI costs helps optimise spend and maximise value per token
Government first
HMRC was the first government and private sector enterprise-level organisation in the UK to enable OpenAI’s GPT 5.2
About HMRC
HMRC (His Majesty’s Revenue and Customs) is the UK’s tax, payments, and customs authority. The Data Platform Services (DPS) division manages HMRCs critical data platforms, providing data access for internal teams and managing massive data clusters, which support policy and operational decision-making.
Industry
Government & Public Sector
Organisation size
67,000 employees
Location
UK
Project length
7 months
Challenge
Bringing complex AI into one of the UK’s largest and most sensitive data environments
HMRC is reforming and modernising the UK’s tax and customs system. So it’s hardly surprising that it has a transformation roadmap that prioritises the adoption of new technologies such as generative AI. However, the importance and sensitivity of data held by HMRC presents a huge challenge when considering the use of public cloud AI services and LLMs.
HMRC’s Data Platform Services (DPS) team also needed to empower civil servants to engage with emerging AI technologies without compromising safety or data sovereignty. Beyond data risk, they also needed to avoid the runaway or unpredictable costs that can come with enterprise-wide LLM access, particularly across a workforce of this scale. Not only that, but the organisation needed to be confident that it could drive adoption at scale by all employees, from frontline staff to case workers and policy specialists. Without this, it would be hard to unlock the full potential of AI to deliver strategic objectives and modernise tax administration.
Solution
Keeping data secure within HMRC’s private internal network
Working in partnership with Equal Experts and third party suppliers, Data Platform Services engineered a bespoke version of OpenAI, that was deployed entirely within HMRC’s secure, multi-cloud network. The Generative AI Landing Zone (GAILZ) is an architecture that ensures no data leaves HMRC’s estate and no vendor has access to HMRC systems or data, eliminating the core risks that prevent most public sector organisations from adopting AI. The innovative multi-cloud environment incorporates both Microsoft Azure and AWS Bedrock, so departments can use the best solution at the best price for them.
To increase organisational adoption, Equal Experts embedded a dedicated product manager who reached beyond Data Platform Services (DPS) into wider HMRC business units to build key stakeholder relationships. This relationship-led approach significantly raised internal awareness of GAILZ, communicating its core value as a secure, approved gateway to LLMs.
The team recognised that early technical constraints created friction during onboarding, so EE designed and implemented ‘Gaitor’ – a purpose-built GAILZ component engineered to simplify the developer and user onboarding experience. By removing these bottlenecks, GAILZ was able to handle the rapid surge in user demand across the department.
GAILZ is designed to offer complete transparency into AI costs and LLM token usage, with multiple spend control layers. Working with EE product managers HMRC technical leadership and 3rd party suppliers, the GAILZ team added features that provide detailed insights into the costs of various models, ensuring stakeholders can use the best tool for each project. By embedding this financial transparency into the product roadmap, it ensured that business leads always retain full control over budget allocation.
The secure LLM gateway also enables access to frontier models, and HMRC secured its position as a global leader in public sector tech by becoming one of the first enterprises in the UK to deploy OpenAI’s GPT 5.2.
Equal Expert’s and HMRC’s dual focus on strategic product management and frictionless technical onboarding helped the platform to scale to 16,000 active users in its very first week. Pairing demand generation with Gaitor’s streamlined onboarding process provided HMRC with an end-to-end operational model that scales safely and remains budget-conscious.
To drive adoption of the new platform, Equal Experts worked with HMRC to build an interactive learning environment that could help build AI skills across the workforce. The Exploration Zone for AI (EZAI) helps staff to use AI safely and effectively to solve business problems, using integrated prompt-comparison tools, a groundbreaking synthetic data library, and gamified compliance challenges. Alongside this, the Rapid AI Innovation (RAIN) apps offers HMRC staff a suite of ready‑to‑use AI utility applications, with productivity gains already being realised in everyday work.
GAILZ has supported civil servants throughout their AI journey, as they move from skepticism or apprehension to full adoption. In addition to the technical delivery, this involved an extensive rollout, including roadshows and demonstrations for all levels, up to the HMRC Executive Committee. This work has generated strong interest from other HMRC teams and external government departments, even being shortlisted for several industry awards.
Results
AI saves workers significant time on utility tasks
The impact of the new tools has been immediate and measurable, positioning HMRC as a leader in secure public sector innovation. The GAILZ project helped DPS to demonstrate how a foundational platform can provide safe and secure access to LLM capabilities, packaging complex AI into accessible, high-productivity tools.
In the first week after its March 2026 launch, the platform onboarded more than 16,000 employees – a quarter of the entire HMRC workforce. Staff are encouraged to identify AI opportunities in their own workstreams, and the AI utility apps mean that some manual processes now take days rather than months to complete.
The integral cost control tools mean that stakeholders have immediate visibility of spending on AI tools, and can easily understand the costs associated with different models for different tasks. Already leading to work to utilise different models now delivering the same business results at a reduced cost.
The hard work behind the scenes has yielded significant results for HMRC, allowing more colleagues to use AI to complete tasks faster, while safeguarding sensitive information, ensuring security and access control through a multi-cloud, multi-vendor approach and – most importantly – providing complete cost transparency. In doing so, HMRC can boost efficiency without compromising on the security standards taxpayers expect.