UK Department for Transport Accelerates Public Policy Insights with Google Cloud AI

Why It Matters

Accelerating feedback analysis lets the DfT meet its 12‑week publication mandate, improves policy responsiveness, and demonstrates how AI can deliver sizable cost savings in the public sector.

Key Takeaways

• DfT saves up to £4 million ($5.1 M) yearly using AI.
• Gemini models achieve 90% accuracy in consultation analysis.
• Tool processes 100,000+ free‑text responses in hours, not months.
• Human‑in‑the‑loop ensures bias checks and policy oversight.
• AI Correspondence Drafter drafts replies using Vertex AI Search.

Pulse Analysis

Public sector agencies have long struggled with the sheer volume of citizen input generated by statutory consultations. In the UK, the Department for Transport receives roughly 55 consultations a year, each producing upwards of 100,000 free‑text comments. Traditional manual coding can take months, delaying the publication of findings and eroding public trust. By deploying Google Cloud’s Vertex AI platform and Gemini large‑language models, the DfT’s new Consultation Analysis Tool (CAT) compresses this timeline to a few hours, while maintaining about 90% thematic accuracy. The speed boost not only meets the department’s statutory 12‑week deadline but also frees analysts to focus on higher‑level interpretation rather than rote categorisation.

The CAT initiative is part of a broader AI ecosystem the DfT is building on Google Cloud. A Connectivity Tool leverages Cloud Run, Cloud CDN and Firestore to model sustainable infrastructure scenarios for urban planners, while an AI Correspondence Drafter uses Vertex AI Search and Gemini to generate first‑draft replies to public inquiries. Crucially, the department adopts a human‑in‑the‑loop governance model, ensuring every AI‑generated insight is vetted for bias, fairness and policy alignment. This responsible‑AI stance mitigates reputational risk and aligns with emerging UK government guidelines on trustworthy AI deployment.

Financially, the automation translates into an estimated £4 million (approximately $5.1 million) annual saving, a figure that underscores AI’s potential to reduce operational costs in government. Beyond the immediate budget impact, the project signals a shift toward data‑driven policymaking, where rapid, evidence‑based decisions can improve transport planning, regulatory adjustments, and citizen engagement. As other ministries observe the DfT’s results, the model may catalyze wider adoption of generative AI across the public sector, accelerating digital transformation while keeping human oversight at the core.