HMCTS uses platform engineering to supercharge digital justice services at scale

1.  Aggregated annually, calculated from peak usage with efficiencies saving each developer 1-hour (or 13.33%) per day
2. Calculated from shrinking lead time to deployment from 120 hours to, typically 4 hours (though often it is as little as 1 hour).
3. Calculated based on increasing deployment frequency from weekly or 0.025 deployments per hour [dph], to teams releasing 10-times per day, or a 1.25dph.

HM Courts & Tribunals Service (HMCTS) is at the heart of the UK’s justice system, responsible for managing criminal, civil, and family courts across England and Wales, as well as a broad range of tribunals throughout the UK. With a workforce of more than 18,000 and a remit that spans everything from serious criminal trials to immigration appeals, family law, and employment hearings, HMCTS plays a critical role in ensuring that justice is accessible, fair, and delivered efficiently.

Traditionally, all casework has been done manually, on paper—a slow and inefficient way to approach modern law. HMCTS embarked on a 10-year modernisation journey to increase capacity and deliver more accessible, efficient, and effective justice for the public. To achieve this, the organisation identified 14 modern digital services to be created to automate time intensive, manual processes and create efficiencies throughout the system. This work represented the most significant digital transformation of the justice system in a lifetime, effectively moving it from paper to digital and into the 21st century. 

Kainos developed a strategy to move HMCTS from its existing ‘DevOps by team model’ to a one focused using platform engineering as an enabler. The vision was to make engineering easier, faster and higher quality by offering a platform made up of APIs, tools, documentation, education and support. Instead of repeating routine infrastructure work for each project, developers access reusable elements via automation and self-service.

The strategy had 5 pillars: 

1. Educate and empower – Readily available documentation, education, tutorials and common solutions to problems 
2. Simplify and evolve architecture – Continually simplifying delivery of solutions to customers, balancing local vs global optimisation 
3. Encourage security and best practice – Security by default and best practice by continuously building and maintaining sensible defaults and codified policies 
4. Self-serve platform – Infrastructure, environments, tooling, documentation and observability available without needing to contact the platform team 
5. Optimise iteration speed, cost and quality – Performance monitored based on DORA metrics and adjusted as teams, systems and the platform itself mature 

‘Golden path’ technologies 

Kainos developed the platform around a standard set of tools and technologies, selected by working with developer teams and choosing the most popular and effective services used across HMCTS.

This provided developers an accessible ‘golden path’ making use of cloud native Microsoft Azure services for networking, containers, storage, caching and databases. The ‘golden path’ also included various management services, including GitHub, CI/CD orchestration and pipelines, security scanning, Infrastructure as Code, Configuration as Code and monitoring.

As a result, developers gained a best-practice, supported way to develop and deploy services with as little friction as possible. Following the golden path makes them fully autonomous for software delivery – from first discovery to go-live, sunsetting and retirement. The fact this golden path removed process friction meant we did not need to mandate adoption across teams, we found teams were asking to be onboarded without needing to ask.

The platform strategy wasn’t about dictating how to do things; it was about empowerment. Kainos therefore implemented it in close collaboration with developers, emphasising the freedom, efficiency and confidence that come from self-service, inbuilt guardrails and the ease of using a platform. Our empowerment approach included conducting ‘Know Your Platform workshops’ with demos and Q&A sessions. We also encouraged developers to innovate off the platform so its ever-maturing offering could evolve alongside the developers and new technologies.

As a result, developer feedback directly influences platform roadmaps using structured processes for gathering this feedback and ensuring a user-centric approach to ongoing platform development.

One of the initial areas identified to improve developer experience was around service desk interaction, with an estimated six thousand tickets created each year. To address this, Kainos developed an AI agent integrated as a ChatOps slackbot to meet developers within their existing slack workflows. The agent automatically integrates with JIRA and uses Azure OpenAI, Azure AI Search, a curated knowledge base stored in CosmosDB alongside retrieval augmented generation (RAG) to provide self-resolution signposting that empowers developers to self-resolve issues.

Since rolling out the Platform Operations service-desk solution, an average 74 percent of interactions have been handled autonomously by the agent assisting developers, with only 13 percent of requests reaching the Platform Operations service desk. By sharing best practices across development teams and using GitHub Copilot to supercharge development, they have seen lead-time for deployment reduced by 3000%, and deployment frequency improved by 4900% using DORA metrics.

Successful digital services delivered using the platform include:

• Reformed digital services for submitting cases
• Documents uploaded to digital cases instead of being posted
• Online civil money claims resolved 3 times faster
• Less than 1% of online divorce applications returned due to user error
• Family public law service application reduced to 9 easy digital screens