Three practices separate organisations that see AI results from the rest
less likely to be right when AI is used on a task it handles badly.
Dell'Acqua et al., HBS, 2023
of top AI performers redesigned their workflows, against 1 in 4 of the rest.
McKinsey, State of AI, 2026. Indicative.
see measurable impact with a clear AI strategy, against 55–60% elsewhere.
BCG, AI at Work, 2026. Indicative.
Your data, and the agents your teams build, stay under your control
We build on whichever models your organisation approves, and we can run every program so that data and agents stay in-country, on infrastructure you control.
Your data stays with you
Exercises run on your approved platforms, or on a sandbox with sample data. Nothing confidential leaves your environment.
Any model you approve
Hosted or open-weight models, whichever your organisation approves, on the platforms you already use.
In-country, on your servers
Open-weight models can run on your own servers or in-country cloud, so the agents your teams build stay sovereign.
How much can AI do before a person signs?
Our framework for deciding how much authority each AI system gets. Select a level to see what it means.
Produces the work; a named person signs before it goes.
Each level is crossed with recoverable or binding consequences. Leaders place their own AI ideas on the ladder, then decide what to fund now, what must exist first, and what moves down a level.
Six commitments apply to every program, whatever its length or sector
Delivered in person by default, in Arabic and English.
Discovery first
Every exercise uses your own problems and processes, on the tools you have approved.
A named output
Every program ends in something your organisation will use. Certificates are awarded only when it is produced.
Governance built in
Every output has a named person who approves it, and is tested for failure before anyone relies on it.
Measured against a baseline
We record where you start. The Impact Report shows what changed, and what to fund next.
Sovereign by design
Your data stays under your control: on platforms you approve, in-country where required, hosted or open-weight.
Reviewed before delivery
Our Advisory Board reviews every program before it reaches an organisation, and members lead selected programs.
Every participant who completes Agentic Build leaves with a demoed agent. If anyone doesn't, we run a follow-up session for the cohort. Terms apply.
Who designs and delivers our programs
Practitioners who build AI design and lead every program. The names and profiles of your program team come with every proposal.
Partners who lead delivery
Our partners design each program and lead its delivery, in Arabic and English. Both have built and run AI programs in the UAE.
A senior Advisory Board
Senior leaders in government, education and space AI review every program before it reaches an organisation, and lead selected programs.
Specialists who have done the work
Function specialists in HR, sales and AI deployment, with a wider network of practitioners contracted per program.
Frequently asked questions
Do participants need a technical background?
No. AI at Work and AI for Your Function need none, and the building programs use no-code platforms, so no programming is required to build a working agent.
Which AI models do you use, and where does our data go?
Whichever your organisation approves: hosted or open-weight, on platforms you already use. Data stays where your policy says. Where no approved platform exists, we supply a sandbox with sample data.
Are programs delivered in Arabic?
Yes. Our team delivers in Arabic and English. Government and education programs are bilingual in materials and delivery.
Where and how are programs delivered?
In person by default, because building on real work needs the room. Online delivery is available for selected formats where it is necessary.
How is impact measured?
Every program records where you start and closes with a four-part Impact Report: baseline, outputs, measured change and next step. The Readiness Snapshot can serve as the baseline.
What happens to my Snapshot answers?
Your result is yours. Your employer sees only group results from 10 or more colleagues, never individual answers. Snapshot data is stored securely by Cader Group.
Can we buy through our existing training provider?
Yes. Our programs are available through training-company partners, co-branded or under the partner's brand, with the same delivery standards and Impact Report.
Will AI replace our staff?
The evidence points to change, not replacement. One in four workers globally is in an occupation with some generative AI exposure, and transformation of jobs is the most likely impact (ILO, 2025). Our programs prepare your people for that change, with a person approving every output.
How does this fit with the AI training our staff already receive?
It builds on it. Awareness courses and e-learning explain what AI is; our programs start there and end in working AI on your own processes, measured against a baseline. Discovery confirms what your teams have already covered, so nothing is repeated.
Where should we start?
Take the free Readiness Snapshot. It recommends a path of up to three steps and ends with the Full Diagnostic or a conversation with our team.
