Agentic AI: From Mandate to Momentum
Enterprises across the UAE are at a defining inflection point. Agentic AI has moved from pilot to national mandate, and the technology stack capable of supporting it, from infrastructure to orchestration to security, is being built in real time.
In this emerging order, the advantage goes to organizations that move fastest from experimentation to execution: deploying agents that don’t just answer questions, but actually get the job done.
The second edition of MIT Sloan Management Review Middle East’s AI Research Forum convenes technology leaders to examine how organizations are actually turning the agentic AI mandate into deployed capability.
Capability that doesn’t just run a pilot, but ships, scales, and holds up in production.
AIRF looks past the adoption headlines to the technology stack that actually determines whether agentic AI scales: infrastructure and compute, orchestration and frameworks, security tooling, and the integration layer connecting agents to the business. It explores what’s actually being built and deployed today, and by whom.
Building on its inaugural edition, AIRF returns to Dubai in 2026 to track how fast the region is actually moving from mandate to momentum.
How can leaders and organizations prepare for an agentic future?
Sponsors
AI Success Partner
Panel Partner
Silver Partner
Past Sponsors
Who is it designed for?
- Business leaders and C-suite executives navigating AI-driven change
- Technology strategists and CTOs seeking to integrate agentic systems
- Policymakers and regulators shaping responsible adoption
- Entrepreneurs and innovators building new products and services using autonomous AI agents
By attending, you will:
- Develop a clear and actionable strategy to integrate Agentic AI into your operations
- Learn how to foster internal collaboration and prepare your teams for AI adoption
- Gain insight into how Agentic AI offers a competitive edge in your industry
Attendees Receive
Livestream access to eight expert sessions
Certificate of attendance
Training sessions with select speakers
Synopsis of each session with key takeaways
Free download of MIT SMR’s exclusive AI whitepaper
Free three-month digital subscription to MIT SMR
I got more out of this four-hour event than three days at a coaching conference.
2023 Summit Attendee
All the speakers were great. I took pages of notes and downloaded all the handouts.
Amazing value! Very interesting presentations and useful insights.
Speakers
We will feature a world-class lineup of thought leaders, innovators, and professors who are actively shaping the evolution of Agentic AI.
To be announced soon
Thought Leaders
Dylan Hadfield-Menell
Assistant Professor,
MIT Computer Science & AI Lab
Tim Kraska
Associate Professor of Electrical Engineering and Computer Science,
Massachusetts Institute of Technology
Marek Kowalkiewicz
Professor and Chair in Digital Economy
QUT Business School
Munther Dahleh
Founding Director of the Institute for Data, Systems, and Society & Professor
Massachusetts Institute of Technology
Industry Experts
Srimanth Rudraraju
Group Director of Engineering
London Stock Exchange Group (LSEG)
Srimanth Rudraraju
Group Director of Engineering
London Stock Exchange Group (LSEG)
Srimanth Rudraraju
Group Director of Engineering
London Stock Exchange Group (LSEG)
Srimanth Rudraraju
Group Director of Engineering
London Stock Exchange Group (LSEG)
A summit covering the essentials of Agentic AI.
Agenda
9:00
Registration & Networking Breakfast
Abbie Lundberg, Editor in Chief, MIT SMR
Elizabeth Heichler, Editorial Director, MIT SMR
Welcoming Remarks from MIT SMR ME
The Readiness Pulse
The AIRF Agentic AI Readiness Pulse: Where the Region Actually Stands
Six months into the UAE’s agentic AI mandate, the AI Research Forum opens with the first independent data on how enterprises are actually responding: where budget is moving and what separates organizations already deploying agents in production.
Panelists:
Geoffrey Alphonso
Chief Executive Officer,
Alef Education
Kaustubh Wagle
Managing Director and Partner,
BCG
Tim Kraska
Associate Professor of Electrical Engineering and Computer Science,
Massachusetts Institute of Technology
Moderator:
Jennifer George
Correspondent,
MIT Sloan Management Review Middle East
Keynote: The Agentic Enterprise: Redesigning Decision Rights for an AI That Doesn't Wait
This keynote examines what agentic AI actually requires of an organization: not an upgrade to existing automation, but a structural redesign of decision rights once systems are authorized to act without human approval at each step.
Rahul Lakhanpal
Vice President of Product Marketing,
DarwinBox
From Mandate to Momentum
Abbie Lundberg, Editor in Chief, MIT SMR
Elizabeth Heichler, Editorial Director, MIT SMR
From Mandate to Momentum: What the Private-Sector Mandate Actually Requires
The UAE’s agentic AI mandate sets specific private-sector obligations against a defined timeline. This session examines what compliance actually requires, how “autonomous execution” is defined in regulatory terms, and the pace at which enterprises are expected to act.
Tim Kraska
Associate Professor of Electrical Engineering and Computer Science,
Massachusetts Institute of Technology
Why Agentic AI Breaks the Existing Technology Stack
Agentic capability can’t simply be bolted onto an existing platform as another feature. This session examines what it actually takes to re-architect enterprise platforms so agents are native to how the system works, not an add-on running alongside it.
Abbie Lundberg, Editor in Chief, MIT SMR
Elizabeth Heichler, Editorial Director, MIT SMR
Panel Discussion: What Would You Hand Over to an Agent Today?
Agentic AI raises a different set of questions once it reaches the top of the organization: not how the system works, but where it actually changes the business. This session examines where agentic AI has moved from a cost line to a genuine source of competitive advantage, what leadership is willing to bet the organization’s direction on, and where the technology still hasn’t earned that bet.
Abbie Lundberg, Editor in Chief, MIT SMR
Elizabeth Heichler, Editorial Director, MIT SMR
12:10
Networking Break
Abbie Lundberg, Editor in Chief, MIT SMR
Elizabeth Heichler, Editorial Director, MIT SMR
From Pilot to Production: What "Two Years" Actually Means for Transformation Budgets
Anirudh Narayan
Co-Founder & Chief Growth Officer,
Lyzr
Panel Discussion: How Enterprises Are Actually Responding to the Two-Year Clock
Panelists:
Awad Ahmed Ali El-Sidiq
Head of Artificial Intelligent & Analytics,
Adnoc Distribution
Srimanth Rudraraju
Group Director of Engineering,
London Stock Exchange Group (LSEG)
Paul Potgieter
Director – Technology and Platforms,
NEOM
Nizar Hneini
Senior Partner,
Roland Berger
Moderator:
Jennifer George
Correspondent,
MIT Sloan Management Review Middle East
Building the Agentic Enterprise
The Infrastructure Behind Autonomous Decision-Making
Autonomous agents that run continuously rather than responding to a single prompt place different demands on infrastructure than conventional AI workloads. This session examines what inference serving, accelerator economics, and state persistence actually look like when an agent’s session can run for hours rather than seconds.
Panel Discussion: When an Agent Makes the Wrong Decision, Who Answers for It?
Autonomy doesn’t remove responsibility, it just makes it harder to locate. This panel examines who actually carries the consequences when an autonomous agent gets it wrong: the team that deployed it, the vendor that built it, or the leadership that signed off on giving it the authority to act, and how accountability frameworks are evolving to keep pace.
Networking lunch
Autonomous agents that run continuously rather than responding to a single prompt place different demands on infrastructure than conventional AI workloads. This session examines what inference serving, accelerator economics, and state persistence actually look like when an agent’s session can run for hours rather than seconds.
Beyond the Chatbot: The Orchestration Layer Agentic AI Actually Requires
The distinction between a conversational assistant and a genuinely agentic system lies in orchestration: the ability to plan, execute, verify, and hand off across multiple steps without intervention. This session examines what that requires in practice, from multi-agent coordination patterns to protocols that let separate agents work together, and where most enterprise deployments are still behind.
Securing an Enterprise That No Longer Waits for Human Approval
Autonomous execution changes the security posture an enterprise needs. This session addresses what it actually takes to treat agents as identities in their own right, with their own permissions and audit trails, and what guardrail and red-teaming practices look like when the primary risk may be an internal agent acting on wrong data.
Keynote: What the Data Says About Organizations That Actually Scale Agents
Survey data consistently shows a wide gap between organizations experimenting with agentic AI and the smaller number scaling it successfully. This keynote examines what the research identifies as the structural differences between the two, and where leadership teams most often make the wrong call.
Governance at Scale: Operationalizing Hundreds of Agents Without Losing Control
Governing a handful of agents and governing several hundred operating across departments are different problems entirely. This session examines what is required to maintain visibility, accountability, and control once agentic AI shifts from pilot to core infrastructure.
Panel Discussion: In the Field, What Early Movers Actually Learned
Production tells a different story than the pilot deck. Leaders who have already deployed agentic AI at scale discuss what broke first once the system went live, where the technology stack did not hold up against real operating conditions, and what they would change with the experience they have now.
Dylan Hadfield-Menell,
Assistant Professor,
MIT Computer Science & AI Lab
The Final Gap: What Still Needs to Happen Before 2028?
With eighteen months left on the clock, the question is no longer whether agentic AI will reshape the enterprise but whether the technology is actually ready to do it at scale. This conversation examines what still needs to be solved at the infrastructure, orchestration, and governance layers before agentic AI can be trusted with the decisions that matter most.
