FEATURED PAPER
By Bob Prieto
Florida, USA
Introduction
This paper looks at disruptive innovation in the project management domain through the combined use of two promising AI technologies. It draws its primary inspiration from DeepMind’s AlphaFold (Appendix 1) success and complements it with the reasoning approach of DeepThink. This combination is focused on providing a foundation for an AlphaFold for Projects disruptive innovation.
In moving through this paper the reader will encounter the following key points:
- AlphaFold for Projects reframes project systems as learnable structures. By treating tasks, assumptions, dependencies, and stakeholders as structured graphs—analogous to residues and contacts in proteins—the approach enables predictive modeling of project behavior using advanced graph‑ and attention‑based architectures.
- DeepMind‑style modeling and DeepThink reasoning form a complementary hybrid. DeepMind contributes the data‑driven, geometry‑aware modeling blueprint, while DeepThink provides the multi‑order reasoning needed to surface assumptions, map entanglement, and interpret systemic dynamics. Together, they create a foundation for a new class of project‑intelligence tools.
- AGI/ADI assumption metrics[1] become powerful model inputs. The Assumption Governance Index (AGI) and Assumption Diffusion Index (ADI) act as “evolutionary signals,” enabling the model to learn which assumptions behave like conserved constraints and which drive volatility across the project system.
- The hybrid model produces calibrated, per‑element confidence for governance. Inspired by AlphaFold’s pLDDT scoring[2], the system generates confidence values for each task and assumption, enabling more consistent governance triggers, earlier escalation, and evidence‑based rebaselining decisions.
- AlphaFold for Projects unlocks new insights into entanglement and propagation. The approach forecasts how local changes—such as assumption migration or resource shifts—propagate through the project network, revealing bottlenecks, systemic vulnerabilities, and high‑leverage intervention points.
- The model enables a suite of high‑value use cases for Large Complex Projects (LCPs). Predictive forecasting, assumption management, bottleneck discovery, intervention ranking, resource optimization, and governance automation become possible through a unified, explainable, confidence‑scored project‑intelligence engine.
In this paper we will describe:
-
- DeepMind and DeepThink
- Relationship between DeepMind and DeepThink
- How they complement each other
- AlphaFold as a relevant analogy
- AlphaFold for Projects
- Its benefits in conjunction with Quantum Project Management[3]
- Potential Use Cases for AlphaFold for Projects (Appendix 2)
DeepMind and DeepThink
DeepThink and DeepMind sound similar, but they sit in completely different categories, and there’s no organizational or historical connection between them. The similarity is only linguistic. Together, they form the foundation for an AlphaFold for Projects concept.
Here’s a clear, structured way to understand the relationship.
More…
To read entire article, click here
How to cite this paper: Prieto, R. (2026). AlphaFold for Projects, PM World Journal, Vol. XV, Issue III, March. Available online at https://pmworldjournal.com/wp-content/uploads/2026/03/pmwj162-Mar2026-Prieto-AlphaFold-for-Projects-featured-paper.pdf
About the Author
Bob Prieto
Chairman & CEO
Strategic Program Management LLC
Jupiter, Florida, USA
Bob Prieto is Chairman & CEO of Strategic Program Management LLC focused on strengthening engineering and construction organizations and improving capital efficiency in large capital construction programs. Previously, Bob was a senior vice president of Fluor, focused on the development, delivery, and turnaround of large, complex projects worldwide across all of the firm’s business lines; and Chairman of Parsons Brinckerhoff, where he led growth initiatives throughout his career with the firm.
Bob’s board level experience includes Parsons Brinckerhoff (Chairman); Cardno (ASX listed; non-executive director); Mott MacDonald (Independent Member of the Shareholders Committee); and Dar al Riyadh Group (current)
Bob consults with owners of large, complex capital asset programs in the development of programmatic delivery strategies encompassing planning, engineering, procurement, construction, financing, and enterprise asset management. He has assisted engineering and construction organizations to improve their strategy and execution and has served as an executive coach to a new CEO. He is author of eleven books, over 1000 papers and National Academy of Construction Executive Insights, and an inventor on 4 issued patents.
Bob’s industry involvement includes the National Academy of Construction and Fellow of the Construction Management Association of America (CMAA). He serves on the New York University Tandon School of Engineering Department of Civil and Urban Engineering Advisory Board and New York University Abu Dhabi Engineering Academic Advisory Council and previously served as a trustee of Polytechnic University. He has served on the Millennium Challenge Corporation Advisory Board and ASCE Industry Leaders Council. He received the ASCE Outstanding Projects and Leaders (OPAL) award in Management (2024). He was appointed as an honorary global advisor for the PM World Journal and Library.
Bob served until 2006 as one of three U.S. presidential appointees to the Asia Pacific Economic Cooperation (APEC) Business Advisory Council (ABAC). He chaired the World Economic Forum’s Engineering & Construction Governors and co-chaired the infrastructure task force in New York after 9/11. He can be contacted at rpstrategic@comcast.net.
To see more works by Bob Prieto, visit his author showcase in the PM World Library at https://pmworldlibrary.net/authors/bob-prieto/
[1] Prieto, R. (2025). Metrics for Assumption Management in Large Complex Projects, PM World Journal, Vol. XIV, Issue XII, December.
[2] The predicted Local Distance Difference Test (pLDDT) score is a measure used in protein structure prediction, particularly by AlphaFold. It assesses the confidence in the local structure with a confidence metric ranging from 0 to 100, indicating the reliability of predicted protein structures, with higher scores reflecting greater confidence.
[3] Prieto, R. (2024). Quantum Project Management, PM World Journal, Vol. XII, Issue I, January 2024.
