Modernising Legacy Systems: Building the Foundations for Public Sector AI Success

Digitisation and legacy modernisation: The foundation for public sector AI success

Digitisation, automation, and updating legacy systems are essential for the public sector to effectively adopt AI. Legacy systems are more than just operational hurdles; they block the path to successful AI implementation.

This issue was a key focus at the recent Think Data for Government event in London, where digital leaders emphasized the need to modernise infrastructure and data foundations. Without this groundwork, AI’s potential remains out of reach.

Chad Bond, director of strategy and innovation at Zaizi, described the challenge as the “great legacy escape.” He pointed out that AI cannot thrive when stuck with outdated systems relying heavily on spreadsheets and inefficient data management. Leaders agreed that simply buying new technology isn’t enough; success depends on addressing core fundamentals.

Richard Appiah, head of data strategy at the Migrations and Borders Group, Home Office, highlighted three essentials: data, people, and culture. Getting these basics right is critical before AI can deliver real value.

Data maturity as a critical starting point

Sian Thomas MBE, chief data officer at the Department for Business and Trade, stressed that infrastructure upgrades must be well justified. If legacy systems cost more to replace than the benefits they provide, organisations face tough decisions.

Dr Ravinder Singh Zandu from the Cabinet Office reminded that AI is a tool, not a magic solution. It only works if the underlying data models are properly trained and the approach to technology adoption is strategic and measured.

Key strategies for effective legacy modernisation

• Start small and demonstrate value
  Richard Appiah advised launching small proof-of-concept projects that quickly show results to senior decision-makers. This gradual approach helps build trust and momentum.
• Develop data literacy
  The Home Office runs “data awareness” events, creating simulated scenarios to highlight the impact of poor data management. This raises understanding across all levels of the organisation.
• Establish clear accountability
  Dr Singh Zandu pointed out the need for clear accountability and a detailed data asset register. Defining responsibilities ensures everyone knows their role in managing data properly.
• Ethical considerations and human oversight
  Sian Thomas emphasised governance frameworks to ensure AI is applied responsibly. “Just because you can do something doesn’t mean you should,” she said. The goal is to develop smarter people and organisations, not just smarter technology. Human insight remains vital in AI deployment.

Practical recommendations

For public sector organisations planning to modernise and prepare for AI, the panel shared these clear steps:

• Conduct a thorough capability assessment
• Develop a defined target operating model
• Create strong data governance frameworks
• Invest in data literacy at all organisational levels
• Start with small, manageable AI pilot projects
• Keep ethics and human oversight front and center

Modernising legacy systems and improving data maturity form the backbone of any successful AI strategy in government. Taking measured, practical steps sets the stage for AI to deliver tangible benefits.