Start here
If you work in the Gulf, you have probably heard that AI "agents" will change how companies and governments run. It is natural to wonder what that means for your team, your business or your job. This section tries to answer that question honestly and calmly.
The short version: organisations are gaining a new kind of tool. It can carry out whole sequences of routine steps, such as reading documents, checking them, updating records and drafting replies, rather than just answering a single question. Used well, these tools free people for the parts of work that need judgment, care and relationships. Used carelessly, they can make mistakes quickly and at scale. Nearly everything useful in this section comes back to one principle: agents can take on tasks, but people keep responsibility.
Why the Gulf is a good place to think about this
The region is not only watching this change. It is actively building for it, and it is doing so with public commitments that put people first:
- On 23 April 2026 the UAE announced a federal framework to move 50% of government sectors, services and operations to agentic AI within two years, with phased rollout, continuous assessment and a commitment to train every federal employee. The announcement ends: "People come first." (Dubai Media Office) [F]
- In June 2026 a single federal Artificial Intelligence and Data Authority was approved to lead data, AI and digital government (Dubai Media Office) [F].
- Gulf regulators are already setting sensible expectations for how AI is supervised. For example, the Central Bank of the UAE's February 2026 guidance asks financial firms to keep meaningful human oversight and a clear ability to stop any AI system (CBUAE) [F].
- A new generation is being prepared: AI is now taught in UAE public schools from kindergarten to Grade 12 [fact base UAE5], and more than six million Saudi public-school students began an AI curriculum in 2025–26 (Saudi Gazette) [F].
The region also knows that building under pressure is real. The 2026 war and the strikes on data centres in the UAE and Bahrain [fact base A1, A3] are a reminder that well-run agentic organisations must be resilient as well as clever.
A short family history of an idea
This section follows one well-known line of thinking about how organisations grow, the Exponential Organizations (ExO) framework, because its newest version speaks directly to AI agents. The framework is not ours. It was created by Salim Ismail and colleagues and is developed today by OpenExO.
| Stage | Year | Core idea | Credit |
|---|---|---|---|
| ExO 1.0 | 2014 | Exponential Organizations: a Massive Transformative Purpose (MTP) plus ten attributes. SCALE (Staff on Demand, Community & Crowd, Algorithms, Leveraged Assets, Engagement) looks outward; IDEAS (Interfaces, Dashboards, Experimentation, Autonomy, Social Technologies) looks inward. | Salim Ismail, Michael S. Malone, Yuri van Geest (PR Newswire) |
| ExO 2.0 | 15 Sep 2023 | Exponential Organizations 2.0: the same structure, updated with eleven attributes ("Algorithms" became "AI & Algorithms"). | Salim Ismail, Peter H. Diamandis, Michael S. Malone (Amazon; OpenExO) |
| ExO 3.0 | 2026 (v25, June 2026) | The Organizational Singularity: the MTP becomes a machine-readable "protocol"; SCALE and IDEAS are replaced by ten characteristics grouped as DRIVE (Decision Architecture, Recursive Learning, Intelligence Stack, Value Moat, Elastic Agency) and SHAPE (Safe Autonomy, Human Architecture, Adaptive Architecture, Purpose Control, Ecosystem Trust). It adds an operating model (the Intelligence Stack with an always-on GOVERN/ASSURE layer) and a six-step playbook (REWRITE). | Salim Ismail with contributors, OpenExO (OpenExO; ExO Model 3.0) |
| Gulf lens | 2026 | The platform's own framing: what agentic organisations mean with Gulf regulation, national workforce programmes, data residency, Arabic-language service and resilience. | The platform (clearly labelled) |
What you will find in this section
- Exponential Organizations explained: MTP, SCALE and IDEAS. Precise, sourced definitions of all eleven ExO 2.0 attributes, with plain-language reflection questions.
- What OpenExO means by ExO 3.0, and a Gulf lens. An attributed summary of DRIVE, SHAPE, the Intelligence Stack and REWRITE, followed by the platform's own view of AI agents as a workforce in the Gulf.
- Anatomy of an agentic organisation. A simple five-part map (purpose, people, the agent loop, data, and an always-on governance layer) linked to public standards and Gulf rules.
- Illustrative archetypes. Five clearly fictional composite organisations, from a 14-person Dubai trader to a Muscat government unit, showing how the ideas might be applied step by step.
Related: the Business section's readiness checklist and digitisation roadmap show how to get from paper to well-governed agents in stages.
How to use this section, whatever your size
- If you run a small business: read the anatomy page and archetype 1. Pick one time-consuming task, write down your "never" rules, and try one assistant with a person approving its work.
- If you lead a larger organisation: read the ExO 3.0 summary, then the governance layer in the anatomy page. Ask: do we have a list of every AI system we use, with a named owner and a way to stop it?
- If you work in government: the Gulf's public commitments give you a clear direction. Focus on transparency, a human route for every citizen, and training.
- If you are a student or early in your career: the skills that grow in value are the ones agents lack: judgment, communication in more than one language, ethics, and understanding how a business really works. Many national programmes now support this learning.
What we will not do
- We will not repeat forecasts as facts. OpenExO's book itself separates its frameworks from its forecasts and calls the forecasts "directional" (OpenExO). Podcast discussions of the framework include large predictions about jobs and company sizes. We do not present these as findings.
- We will not present fictional examples as real. Our archetypes are labelled illustrative on every page.
- We will not claim OpenExO's work as ours. Where we add something, we say so.
A note on permissions
ExO, the ExO attribute names, the ExO Mind Map™, ExO 3.0, DRIVE, SHAPE, the Intelligence Stack, GOVERN/ASSURE, REWRITE and The Organizational Singularity belong to their authors and to OpenExO. We summarise them for education, with links, and we do not reproduce their diagrams or long passages. Before publishing this section, the platform should ask OpenExO whether using these names, and any diagrams derived from them, needs permission or a licence, particularly as the platform plans a membership model and a self-assessment tool.
Join the conversation
Questions, experiences and respectful disagreement are welcome in the forum's Agentic Organisations & ExO space (F10). If you spot an error on these pages, please report it in Site Feedback & Corrections (F18). We publish corrections openly.