INSEAD · Technology & Operations Management

Michael Freeman

As an INSEAD professor, I work between rigorous research and the real choices leaders face. Most recently, that has meant designing agentic AI you can trust, through a method of my own, Design the Agent.

01 About

Empirical evidence on the decisions that run organisations and the design of systems that improve them.

Michael Freeman
Associate ProfessorSingapore

Michael Freeman is a tenured Associate Professor of Technology and Operations Management at INSEAD, and holder of its Impact Fellowship. He earned his PhD in Management Science from the University of Cambridge (Judge Business School), where he remains a Research Fellow at the Cambridge Centre for Health Leadership & Enterprise.

His award-winning research, with six papers in Management Science, has reached the front pages of national newspapers. Today his work turns on a question every executive team faces: how people and algorithms should share decisions, and how to design AI agents an organisation can trust.

He teaches across INSEAD's MBA, Global EMBA, PhD and Executive programmes, including the flagship Advanced Management Programme for C-suite executives, and has twice received the school's Executive Education Award for Outstanding Teaching.

02 Executive Education

Building the future-ready organisation.

Michael works with executive teams to address real business challenges and uncover opportunities they can act on: strategic reinvention, operational and cultural transformation, and effective AI strategy. He has partnered with senior leaders across industries, from healthcare to financial services, in bespoke company engagements.

INSEAD · Michael Freeman on innovation & change
/01

Strategy & AI

Reading the signals that matter, including what AI makes possible and what it changes, building strategic foresight, and translating what's coming into a plan the organisation can act on.

/02

Innovation & disruption

New business models, the dynamics of disruption, and growth strategies that hold up when the ground is shifting: how to read where value is moving and build the models to capture it.

/03

Operational transformation

Turning strategy into how the organisation actually runs: redesigning processes, rewiring culture, and building the operating model that makes change stick.

Formats
  • Keynote
  • Workshop
  • Multi-day programme
  • Week-long programme

From a single keynote to a week-long intervention on organisational and cultural transformation, built bespoke for your company.

Signature sessionFrom Chatbot to Agentic AI

A hands-on session on how modern AI agents really work, beyond the chatbot. Leaders look inside a real agent, then work through the design decisions behind whether one can be trusted once it is live. Based on the Design the Agent methodology and the forthcoming book.

How the method works →Enquire about the AI session →
Signature sessionThe Phoenix Encounter

An INSEAD method for radical reinvention. The leadership team imagines its business burned to the ground by disruption, then rebuilds it to rise stronger from the ashes. A structured way to confront what could destroy you, and move first.

The method, in MIT Sloan Management Review →Enquire about the Encounter →

Selected organisations Michael has taught and advised

Sony GroupSMBCStandard CharteredKPMGProcter & GambleAccentureMedtronicLenovoColgateNTT DATATata Consultancy ServicesJardine MathesonOCBCSonepar
InfineonKyowa KirinJanssenSumitomo ChemicalFairPrice GroupRoyal Golden EagleMedia PrimaBanglalinkSakata SeedJubilant PharmovaISB AlumniGC ChemicalsHealthcare Leadership College

Company-specific and customised executive programmes, 2019–2025.

03 Flagship work

Don't prompt.
Design.

Most agentic AI systems fail for a reason that gets worse as the models improve. Every design decision you skip before deploying an agent is a debt, and model capability is the interest rate.

Design the Agent is Michael's framework for paying that debt down while it is still cheap. It centres on the Agentic AI Design Canvas, a one-page tool for the nine decisions that determine whether an agent can be trusted, and on the Agent Operating Model, which sees every agent as three layers: the worker that thinks, the harness that controls, and the tools that give it reach.

The Agentic AI Design Canvas: a one-page design tool with nine decision cells in three groups.
The Agentic AI Design Canvas · CC BY-SA 4.0

04 The book

Michael Freeman Don't Prompt,
Design
The discipline of agentic AI design

The method, in long form.

Don't Prompt, Design: a field guide to designing agentic AI, for the executives who have to answer for it.

It turns that method into a working playbook: the nine decisions, the failure modes, and real cases of agentic AI succeeding and failing, drawn from the work Michael does with executives at INSEAD. It is written for leaders who are deploying agents faster than they are designing them.

In development

I'll let you know when the book is ready, and send the occasional article or early chapter in the meantime. Unsubscribe anytime.

05 Research

How operational choices and new technologies shape the service people get.

Three threads run through the work: how the design of service operations shapes outcomes, how organisations decide under uncertainty, and, most recently, how human and machine intelligence can work together.

/01

Healthcare Operations

How the decisions embedded in healthcare delivery shape patient outcomes at scale, and what rigorous empirical evidence can tell organisations about designing it better.

/02

Empirical Operations & Decision Quality

Large-scale causal analysis of how organisations make operational decisions inside multi-stage systems, drawing on datasets of millions of records.

/03

AI & Algorithms in Operations

How people and algorithms share the work: algorithmic gatekeeping, AI guidance in tiered services, and the design of agents that can be trusted.

Selected publications

  1. Kajaria-Montag, H., Freeman, M., & Scholtes, S. (2024). Continuity of care increases physician productivity in primary care. Management Science, 70(11).

    Front-page national coveragedoi.org →

  2. Freeman, M., Robinson, S., & Scholtes, S. (2021). Gatekeeping, fast and slow: An empirical study of referral errors in the emergency department. Management Science, 67(7).

    doi.org →

  3. Freeman, M., Savva, N., & Scholtes, S. (2021). Economies of scale and scope in hospitals: An empirical study of volume spillovers. Management Science, 67(2).

    First Prize, MSOM Student Paper Competitiondoi.org →

  4. Freeman, M., & Ding, J. (2026). Algorithmic gatekeeping. Under review, Management Science.

    SSRN →

View all publications → Also on Google Scholar.

06 In the media

Research that reached the front page.

A study Michael co-authored on continuity of care was front-page news in The Guardian and The Telegraph on 23 February 2024, and was also covered by the BBC, The Times, The Independent, and across the trade press.

07 Contact

Get in touch.

For executive-education enquiries, speaking, media, or research collaborations, send a note and I'll get back to you. You can also email directly.

michael.freeman@insead.edu

INSEAD · 1 Ayer Rajah Avenue · Singapore 138676

LinkedIn → Google Scholar →