Mott MacDonald Deploys Over 8,000 Copilot AI Agents Across Engineering Operations

Mott MacDonald staff have built more than 8,000 AI agents using Microsoft Copilot, according to global head of AI and data Nasrine Tomasi in an interview with Microsoft. The software deployment expanded from an initial 50 licenses to 1,000, then 5,000, and is now rolling out to all 20,000 employees. Engineering use cases include screening approximately 220,000 requirements on an airport project to highlight around 200 conflicts, and structural analysis on an underground station that showed temporary works could safely serve as permanent structure. In bidding, a hydropower team uses agents to assemble requirements and calculate budgets, aiming to reduce proposal preparation time from one week to 48 hours. Rapid proliferation has introduced governance issues including duplicate solutions, dormant tools, and knowledge gaps when creators depart. In response, the firm is building a central AI library and expanding its internal assistant EMMA into structured project review workflows while preparing Microsoft Fabric as a central data platform.

Key takeaways

• Staff have developed over 8,000 AI agents using Microsoft Copilot, with license deployment expanding toward all 20,000 employees.
• Engineering applications include analyzing 220,000 airport project requirements to flag 200 conflicts and verifying structural reuse of temporary works on an underground station.
• Proposal teams use agents for bidding tasks to cut proposal cycles toward 48 hours, while internal assistant EMMA captures structured insights across monthly project reviews.
• To address overlapping tools and dormant agents, Mott MacDonald is deploying a central AI library and preparing Microsoft Fabric as a unified data platform.