Case Study:

AI-Powered QA for Government Data Digitisation

Our Client

A major UK government body responsible for maintaining the national archive of civil registration records – including births, deaths, marriages, adoptions, and more. With over 280 million records dating back to the 18th century, it’s one of the country’s most vital sources of legal and historical data.

The challenge Our role Our approach The solution Outcomes

The challenge

A major digitisation programme was underway to preserve aging paper and microfilm records and improve public access. But the scale was immense: 200,000 records needed to be transcribed per day over a 2–3 year period.

With that volume, reviewing each record manually wasn’t viable. The risk? That millions of records would go unchecked, inaccuracies would slip through, and the data could lose its integrity – risking public trust and future use in services like identity verification.

Our Role

We were brought in to explore how AI could support the quality assurance process. The goal wasn’t to replace human reviewers, but to help them focus their time where it mattered most.

Our Approach

We developed two Proof of Concepts (PoCs):

One to assess the quality of scanned record images

One to review transcribed data for potential errors

We used advanced computer vision techniques (CNNs) and synthetic datasets to train models that could flag low-quality scans and possible transcription issues. This let human QA teams focus only on the records most likely to contain errors.

Working closely with stakeholders across departments, we tested our approach on live data. The models achieved accuracy rates between 80–95%.

The solution: AQuA (Automated Quality Assurance)

Following successful PoCs, we were commissioned to build the full solution: AQuA.

AQuA works in two ways:

It checks records in real time as they're transcribed, flagging those that need human review.

It reviews all records after handover to ensure supplier KPIs are being met and data quality holds up.

AQuA has reduced the need for random dip sampling, improved team efficiency, and significantly increased overall record quality.

Outcomes

Records flagged by AQuA are more likely to contain real issues, allowing targeted reviews

QA teams now work on fixing, not just searching

Supplier performance is easier to track and challenge

The tool has potential to be used across other public sector programmes that require large-scale data and image validation

Conclusion

This project showed what’s possible when innovation is applied pragmatically. AQuA didn’t just introduce advanced technology – it delivered measurable, ongoing impact.

By combining AI with human expertise, Rising Tide AI created a solution that makes a critical national service more accurate, efficient, and future-ready. The result is a tool that builds trust, improves public service delivery, and proves how thoughtful automation can scale across government – without replacing the people who keep it running.