Rewiring the State — Eoin Mulgrew, No. 10 (Downing Street)

2026-05-18 · Source: AI Engineer · Field: Government & Public Sector — Digital Government & E-Government, Public Policy & Governance, Public Safety & Security · Depth: Intermediate, extended

Summary

The 10 Downing Street Data Science team (10 DS), established during the pandemic, is scaling its AI engineering and development capabilities to inform critical government decisions and drive AI adoption across the UK state. The team addresses a significant public sector productivity crisis, evidenced by 7.25 million people on NHS waiting lists and 350,000 court case backlogs, with an estimated £40 billion annual productivity prize from AI in government. Facing challenges like uncompetitive pay and bureaucracy, 10 DS employs an "insurgency model" with high political backing, market-rate pay, and autonomous, technically focused recruitment (0.7-0.8% success rate) to attract external talent. This model enables rapid development of internal AI tools, such as policy simulation and statute book analysis, and fosters partnerships with entities like the AI Safety Institute, Incubator for AI, and Just AI to deploy solutions like the Extract tool for planning applications and AI tutors for education.

Key takeaway

For government executives and technical leaders aiming to modernize public services, your focus should be on empowering small, agile technical teams with direct political support and the autonomy to recruit top external talent. This approach, exemplified by the 10 DS "insurgency model," can bypass traditional bureaucratic hurdles, rapidly deliver impactful AI solutions, and foster an ecosystem of innovation that addresses critical productivity gaps and improves service delivery at an unprecedented pace.

Key insights

Small, empowered technical teams with political backing can rapidly deploy AI solutions to address public sector challenges.

Principles

• Political will enables rapid tech adoption.
• External talent can drive internal transformation.
• Focus on high-impact, low-hanging fruit first.

Method

The "insurgency model" involves a small, central team with a mandate from top leadership, market-rate pay, high autonomy, and a rigorous, technically focused recruitment process for external experts to rapidly implement AI solutions.

In practice

• Implement policy simulation tools for faster, data-informed decisions.
• Automate legal analysis to save costs and increase speed.
• Develop public-facing dashboards to enhance transparency.

Topics

• No. 10 Data Science Team
• Government AI Adoption
• Insurgency Model
• Forward-Deployed Engineering
• AI Safety Institute

Best for: Executive, AI Engineer, Director of AI/ML, Policy Maker

Related on AIssential

• Government adoption of AI agents could outpace the private sector — Agentic AI is a government mandate, driving operational transformation and hybrid human-AI workforces.
• Teachers aren’t getting formal guidance on AI, poll finds — Rapid AI evolution and escalating global conflicts demand integrated governance and adaptive strategies to manage both technological and geopolitical risks.
• $500 million AI jobs push launches with bipartisan backing - Politico — A bipartisan private-sector initiative addresses AI job displacement through worker retraining and policy innovation.
• Building a mission-led government: Three essential actions for delivering lasting change — Governments must reclaim digital core sovereignty, enable composability, and leverage AI to achieve mission-led, citizen-focused outcomes.
• Stop Shouting. Start Policymaking. — AI-driven platforms can identify hidden public consensus, transforming divisive political discourse into actionable policy.
• Why government agencies need structural modernization before AI adoption — Successful government AI adoption requires structural modernization and robust engineering foundations to balance speed with public trust.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.