Yoshi Soornack

Associate Director, AI & Data,Gleeds LLP

Yoshi Soornack is Associate Director of AI & Data at Gleeds, where he leads the firm's AI developments and drives transformation across project delivery. As a former neuroscientist with published research in eye-tracking and brain stimulation, he brings a distinctive perspective to the intersection of cognitive science and artificial intelligence in the built environment.

During his three years at Gleeds, Yoshi has also served as Digital Community of Practice Lead at CASA (formerly ACOSTE), and prior to this he managed the UK's largest project data analytics community. He was a founding member of the Project Data Analytics Task Force, which significantly influenced the Infrastructure and Projects Authority's landmark report, Data Analytics and AI in Government Project Delivery (2024).

Co-hosting Project Flux, a pioneering podcast and newsletter exploring AI's impact on construction and project delivery, Yoshi educates and inspires industry professionals to embrace AI for more efficient, adaptable delivery. Key interests are systems and emergent complexity.

Session:  Closing the gap - A practical framework for implementing Data Analytics in the project environment' session.

Most enterprise AI pilots fail to deliver measurable return. MIT's recent research shows that approximately 95% of GenAI investments never make it from experiment to production. Yet at Gleeds, we've built an AI capability that's delivering 140+ hours saved per project, transforming delivery timelines from days to minutes, and embedding AI into the daily workflows of over 1,200 construction professionals.

This session reveals why most implementations fail and how we succeeded by navigating the goldilocks zone between oversimplification (shelfware that gets ignored) and overcomplication (governance committees that never ship). We'll share the hard-won lessons from scaling AI across a complex, multi-disciplinary construction consultancy - from cost management and project planning to risk assessment and commercial strategy.

You'll learn:

• Why the AI hype cycle softened and what MIT's research reveals about the implementation trap
• How to apply Ashby's Law and Gall's Law to avoid the extremes that kill AI programmes
• Our lean scaling model: hub-and-spoke champions, use-case matrix (complexity vs value), and workflow embedding
• The 8 target behaviours that differentiate superficial tool usage from genuine capability integration (auditability, safety, verification, adaptability, collaboration, reusability, decision-support, real-world application)
• Real case studies from education, infrastructure, and project management showing 7-140 hour time savings and better decision quality
• The transformation cycle that takes AI from workshop probes to embedded standard operating procedures
• Before/after scenarios showing how AI-ready delivery transforms Monday to Friday workflows - from service execution plans and risk workshops to change orders and monthly reporting

This isn't theory. We'll show you the compounding efficiency gains across real PM workflows, the upskilling approach that bridges surface understanding to deep operational mastery, and how we aligned with the RICS Global Standard (effective March 2026) and APM Body of Knowledge 8th Edition requirements for responsible AI use.

Whether you're early in your AI journey or struggling to move pilots into production, this session offers a practical blueprint for escaping the 95% failure rate and joining the 5% club that actually delivers measurable value.