Reflections from Modernising Criminal Justice 2026
Data is one of the most important disciplines in public services, yet it doesn’t often receive the attention it deserves until it is urgent. That observation, from Sumitra Varma, Deputy Director of Probation and Electronic Monitoring Data at the Ministry of Justice, set the tone for Version 1’s fireside chat at the Modernising Criminal Justice conference in June 2026.
About the event and key themes of the day
The Modernising Criminal Justice Conference brings together organisations that make up the UK justice system function: the Ministry of Justice and the agencies within the department, local police forces, courts, tribunals, and training and skills bodies, among others. What makes it useful for attendees is its narrow focus, bringing together leaders dedicated to delivering a world class justice system. Frontline staff sat alongside Chief Digital Officers, judges and ministers, which meant discussion stayed grounded in operational reality rather than drifting into the abstract.
Throughout the day it was clear that the outcomes that matter most to the public and the taxpayer are the outcomes that should set the agenda. Speakers returned repeatedly to four priorities:
- Speeding up the court system
- Reducing reoffending rates
- Improving outcomes across the justice system
- Applying digital and artificial intelligence in ways that demonstrably improve those outcomes
Programmes that connect their work to one or more of these measures have a clear case for continued investment. Programmes that can’t make that connection face harder questions.
The Conversation around Data has changed
What stood out across the sessions was the maturity of the conversation about data. The debate is no longer about whether data matters. Instead, attention has shifted to the practical barriers that prevent organisations from using data effectively.
Success now depends on three things:
Treating data as a strategic leadership priority
Enabling secure and lawful data sharing across organisations
Improving data quality so that operational decisions can be made with confidence
As interest in AI grows, these foundations become increasingly important. High-quality data is no longer simply desirable; it is essential.
Our Session with the Ministry of Justice
We hosted a fireside chat with Sumitra Varma, Deputy Director of Probation and Electronic Monitoring Data at the Ministry of Justice. Sumitra leads the data science, data engineering and data quality management capability behind some of the most consequential decisions in the justice system, from frontline risk assessment through to electronic monitoring and interventions to reduce reoffending.
Probation sits at the sharp end of criminal justice reform. With prisons under sustained pressure, the probation service carries growing responsibility for keeping communities safe while supporting people to rebuild their lives. Getting that balance right depends on the right information reaching the right hands at the right moment.
The framing for the conversation was direct: data is one of the most important disciplines in public services, and while it often doesn’t receive the attention it deserves, at the Ministry of Justice, this is changing. The appetite for artificial intelligence and quality insights makes the data gap impossible to ignore, because the quality of the data sets the ceiling for the quality of the AI built on top of it.
Sumitra was joined by Dimitris Perdikou, Digital Advisor at Version 1, and Brian Hodgson, Justice Sector Lead at Version 1 who chaired the discussion.
The data exists, practitioners need to access it
Sumitra opened the discussion by describing a live problem her team is working on: understanding what data is captured in systems today, what data a probation practitioner actually needs to do their job well, and how to ensure they have ready access to that data at the moment they need it. When a probation officer sits down with someone on their caseload, the quality of that conversation depends on the picture available to them in the room. The Ministry of Justice is investing in improving data availability, accessibility and quality to help practitioners make the most of those interactions and support better outcomes for people on probation.
Dimitris set out three common barriers for this gap between data and using it in practice, across most government departments. The first is data ownership, because the data that informs a practitioner sits with different owners both inside and outside the organisation. The second is connecting data across environments, where legacy systems sitting in ageing data centres were not built to connect to modern tools, and retrofitting them is slow and costly. The third is timeliness: the need for near real-time access to information during frontline interactions.
The conversation also returned to the value of digital professionals spending time in operational roles. Sumitra has been driving frontline visits across her team, including visits to court, open prison and probation centres. The practice surfaces day-to-day realities and constraints that no system specification ever captures on its own.
Reducing reoffending through targeted intervention
Reducing reoffending remains one of the justice system’s most significant challenges.
The evidence on reoffending points consistently towards relationships and community connection as protective factors. The challenge is to build a data environment that supports that kind of work rather than reducing it to compliance monitoring.
Sumitra’s team is working with policy and analysis colleagues on more targeted interventions and pilots, with the aim of moving beyond broad approaches towards precise, evidence-based action. The goal is to move beyond broad approaches and identify what works for specific groups and circumstances.
Dimitris reinforced the value of international evidence, noting that research from comparable jurisdictions around the world is invaluable input to policy decisions in the United Kingdom. Norway, for example, reduced its reoffending rate from around 60–70% before its 1990s prison reforms to 20% today. He also argued for openness: where the data and the decisions can be made public, doing so helps bring the public alongside the choices being made, which is itself part of the reform.
Tools inform, humans decide
Risk assessment tools have long supported probation practice, but their effectiveness depends on the quality and completeness of the underlying data. Actuarial risk scoring has been part of probation practice for years, but as with any data tool, its accuracy is limited by the cross-government data available to feed it. Technology should support professional judgement, not replace it. Systems should be transparent, easy to use, and capable of presenting relevant information in a way that helps practitioners understand why a risk assessment has been generated.
The opportunity discussed on stage was to combine improved scoring with practitioner judgement, not to substitute one for the other.
Dimitris set out a clear position on how the systems should behave. They should be straightforward for the practitioner to use, transparent about where risk analysis is being applied, and configured to allow drill-down at the point of decision. If a person is being risk-assessed because of the nature of the offences they have committed, the system should provide a quick summary of those offences. If risk relates to a health issue, the system should show that.
The principle that ran through the discussion was that the human takes the decision, and the system’s job is to put as much relevant information at their fingertips as possible, in a form that supports rather than displaces judgement.
Unlocking the potential of Electronic Monitoring
The Ministry of Justice is actively looking into opportunities to use electronic monitoring. While location monitoring is often associated with compliance, its potential extends far beyond that.
For example, in one of the pilots being done, if someone who has previously been convicted of acquisitive crime (burglary, theft etc) on electronic monitoring is in the community on probation at the time of a local incident, their location data can place them elsewhere and rule them out entirely, saving valuable police time. Sumitra’s team is already exploring multiple data science and AI opportunities within Electronic Monitoring, moving beyond basic compliance tracking towards using monitoring data to support active decision-making. These pilots not only help to improve public protection but also serve to reduce reoffending rates while saving valuable police time.
Dimitris widened the lens to geospatial data more broadly, it’s a fast-growing area across government. The National Data Library is already using geospatial data to support the utilities industry and is extending into other areas. The argument he made was for opening that data to everyone involved in the justice system, with emergency services having it to hand during incidents.
The caution he attached was equally important: the framing must maintain public trust by clearly communicating the benefits and safeguards associated with these technologies.
AI is advancing but data foundations must come first Artificial intelligence was a natural landing point for the conversation, given the framing at the start. The position from both speakers was consistent: data quality is the prerequisite for AI that works effectively in justice settings, and shortcuts here produce systems that fail in ways that are difficult to detect and costly to correct.
Dimitris noted that what is genuinely possible today was not possible two years ago, which sharpens the case for getting the data foundation right now. The session closed on what needs to change.
While the Ministry of Justice is focussing on strengthening its data foundations, Sumitra was clear on two systemic challenges. Data sharing with other government departments needs to be easier, and data quality captured via digital systems used by frontline officers needs to constantly improve.
Dimitris added a third thread: the value of bringing policy teams closer to early iterations of new ideas, on the basis that modern artificial intelligence tools are making that kind of co-development more achievable than it has been historically.
If those three threads can be pulled on together, the foundation for the next phase of justice reform looks considerably stronger.
If your organisation is working on a justice modernisation programme, or on the data foundations that sit beneath one, we are offering sessions at our AI Co-Creation Studio, where we work alongside public sector teams to develop practical, governed AI solutions grounded in real operational problems.
To find out more or to book a session, contact:
