SAUDI TOOK AI TO EIGHT LOCATIONS.

WHAT HAPPENED

Monsha'at, Saudi Arabia's SME authority, and SDAIA, its data and AI authority, ran an AI Week from 9 to 13 Aug 2026 through eight SME support centers — Riyadh, Jeddah, Al-Khobar, Madinah, Aseer, Jazan, Qassim and Al-Baha. Aseer and Qassim are regions rather than cities. Reported outcomes: more than 3,900 people through the program, over 40 public, private and educational bodies involved, 17 exhibitors, six main sessions and 78 entrepreneurial meetings. It sits inside Saudi Arabia's designation of 2026 as its Year of Artificial Intelligence.

THE TRAINING WENT TO SMEs, NOT ENTERPRISES

Not enterprises. Not government departments. Small and medium businesses, in eight locations, several of which are nowhere near the capital.

That is a deliberate bet, and it is a different one from the region's headline AI spending. Data centers, sovereign compute and national models are supply-side plays: they make capability available. Taking a week of workshops to Jazan and Al-Baha is a demand-side play: it makes capability usable by the companies that employ most people and have no AI function at all.

Both are needed and they fail differently. Oversupply without demand leaves you renting capacity to foreign customers. Demand without supply leaves you dependent on someone else's infrastructure. Watching which side a government is currently funding tells you where its own diagnosis says the bottleneck is.

YOUR 2028 SAUDI COUNTERPARTIES ARE IN TRAINING NOW

Two reasons, one competitive and one practical.

The competitive one: the state-funded track behind this week is the larger fact. SAMAI has reported over one million participants and 11,000 trained specialists (Arab News, 11 Mar 2026), and AI Week is that program's SME-facing edge rather than a one-off. Over two or three years that compounds into a supplier base that quotes faster and costs less to work with. If your growth plan includes the Kingdom, the counterparties you will be negotiating with in 2028 are being trained now.

The practical one: it is a free curriculum design. The program's structure — short sessions, mixed audience, exhibitors alongside training, meetings attached — is a reasonable template for internal enablement. Most training programs we are asked to review are a single session built on a vendor's demo data. What changes behavior is what this program used: repetition on the participant's own cases, with someone in the room who can answer "how would that work in our process".

There is a second-order effect worth watching. Programs like this generate a supply of people who have used the tools on real work, and those people become hires. A UAE company recruiting for an operations or analyst role in 2027 will be reading CVs from a candidate pool that has been through state-funded AI training, and the ones who have not will look expensive to onboard by comparison. Enablement policy is quietly a labor-market policy.

TRAINING PEOPLE COSTS LESS THAN BUYING SOFTWARE

• Enablement is cheaper than procurement, and usually removes the need for it. The constraint on AI adoption in a mid-sized company is almost never licenses. It is the number of people who can specify a workflow precisely enough to automate it.
• Go where the work is. Training that happens in the department, on that department's real cases, converts. Training that happens in a conference room on generic examples does not.
• State-funded literacy is a competitive input, not a courtesy. Ask Saudi suppliers what their staff were trained on. It is now a normal qualification question.
• Measure attendance against changed workflows, not against headcount trained. 3,900 participants is an input. The output is how many of them shipped something.