Gen AI and legacy IT: A public sector guide
Generative AI has shone a spotlight on barriers presented by legacy IT. With growing opportunities for savings and value creation, the public sector needs ways to harness it securely, ethically, and efficiently.
Importantly, efforts to tackle legacy IT don’t need to involve ‘big bang’ digital transformation. There are straightforward, iterative steps that deliver short-term gains and long-term value.
Let’s look at these opportunities, including:
-
How to balance effort, risk and return in identifying generative AI projects
-
Legacy data barriers to overcome
-
Don’ts and Do’s for efficient, cost-effective data transformation
Balancing effort, risk and return with generative AI
Generative AI isn’t ‘all or nothing’. Different adoption levels require different amounts of effort and deliver different returns.
So where should UK public sector organisations start, given that effort/risk/return equation?
Prioritising public sector use cases
You can gain lower-level benefits almost immediately with less complicated use cases, like internal meeting summarisation and citizen engagement transcription and translation.
Other use cases with higher returns in terms of efficiency gains and citizen experience improvements generally require greater effort and investment in data platforms. These include:
-
Operational efficiency and resource management – automated document processing / data analytics for performance insights / fleet management optimisation / claims processing automation
-
Financial management – debt collection and fraud detection / benefit and council tax assessments
-
Citizen service and engagement – digital assistants / email and call management / accessibility and inclusivity by design
The legacy data barrier – Data Quality Matters
However, to get started with these use cases, you need to tackle certain legacy IT barriers, specifically on the data side. After all, generative AI output quality is only as good as the data foundation.
In most public sector organisations, data is held in disparate locations, multiple formats, and many technologies. The inability to extract useable data has been cited as one of the greatest barriers to process modernisation and innovation across government.
So how do you start consolidating and transforming that legacy data estate?
Don’ts and Do’s for an efficient, cost-effective data transformation
Let’s start with the Don’ts:
Don’t think of it as a Pandora’s Box. No matter how fragmented and complex the legacy data estate is, you can make progress without untangling everything at once.
At the same time, don’t ignore legacy systems deemed to be ‘working okay’. Often, these only get enough investment to keep them running. This isn’t sustainable – you risk creating much more complex, expensive challenges down the road.
Now, let’s look at the Do’s:
Think about incremental wins. The goal is to continuously improve while safeguarding ‘business as usual’ operations and meeting user needs.
Best-practice steps for this incremental legacy transformation are:
-
Focus on users to minimise disruption – build support with users and enable effective change activities
-
Take a front-end-led approach – making back-end changes without impacting users
-
Identify opportunities for change – automate manual tasks and remove redundant processes
-
Analyse legacy – including dependencies and requirements
-
Highlight dependencies – to support longer lead times for change
-
Clearly articulate boundaries and dependencies – between legacy and new
You also need to embed the appropriate data strategy, encompassing governance, operations, and culture. This is essential for compliance and security. It also drives incremental transformation by protecting against a reversion to old approaches. To achieve this, you need to address 6 key areas:
-
Data transformation vision and strategy – what the organisation is trying to achieve with data
-
Governance approach – how the organisation defines and enforces data policies, controls, processes and standards
-
Information operations – management of data in the end-to-end lifecycle, including how, when and by whom it’s created
-
Organisational culture – the structures, roles, and accountability to manage data and enable it to be a collective responsibility
-
Insights and visualisation – defining use cases that drive value from the data, as well as measuring and demonstrating governance progress and its impact
-
Enterprise architecture – the ‘master plan’ for your data-driven technology, guiding development by prescribing goals and requirements
A legacy data estate shouldn’t hold you back
Barriers from a legacy data estate are manageable (and addressing them is more straightforward than you might think). Just keep these two points in mind:
-
Plan an incremental transformation – so you embed best-practice data governance and drive continuous change with longer-term benefits
-
Leverage proven solutions instead of building from scratch – solutions like data platform as a service not only save time and budget, but streamline security and compliance
With these ideas underpinning your approach, you can hit the ground running with generative AI, achieving more for less across operations and citizen services.
Heather Cover-Kus
Heather is Head of Central Government Programme at techUK, working to represent the supplier community of tech products and services to Central Government.
Ellie Huckle
Ellie joined techUK in March 2018 as a Programme Assistant to the Public Sector team and now works as a Programme Manager for the Central Government Programme.
Annie Collings
Annie is the Programme Manager for Cyber Resilience at techUK. She first joined as the Programme Manager for Cyber Security and Central Government in September 2023.
Austin Earl
Austin joined techUK’s Central Government team in March 2024 to launch a workstream within Education and EdTech.
Ella Gago-Brookes
Ella joined techUK in November 2023 as a Markets Team Assistant, supporting the Justice and Emergency Services, Central Government and Financial Services Programmes.
Authors
