Plaith
Plaith
Toggle theme

Operational Systems Studio · Dubai, UAE

Maturation curveMay 12, 2026

Operational AI is in 2005

Web development was once stuck in a phase where the technology worked but the products that operated it at scale hadn't been built yet. That phase was 2005. Operational AI is in that phase right now.

Web development was once stuck in a phase where the technology worked but the products that operated it at scale hadn't been built yet. That phase was 2005. Cloud went through it in 2012. Mobile around 2010.

Operational AI is in that phase right now.

The model layer is in extraordinary shape. ChatGPT, Claude, Gemini, and the open-source Chinese frontier from DeepSeek ship multimodal, state-of-the-art capabilities every few weeks. Closer to home, regional models like Jais 2 run on sovereign clouds in Arabic. The model layer has never had this much capability moving through it.

What's missing is the layer between those models and a working consumer brand operation. The orchestration that runs multi-brand campaigns across UAE, Saudi, and Qatar without somebody manually rebuilding the workflow each quarter. The pipeline that turns Arabic and English content production into a continuous output rather than a project commissioned from scratch each Ramadan. The system that absorbs the regional approval cycles quietly consuming marketing teams across the GCC.

Per Deloitte's State of AI in the Middle East, 80% of regional organisations feel pressure to adopt AI and 69% plan increased investment, with UAE working-age AI usage now at 64% — the highest in the world. The demand is there. The infrastructure to convert that demand into operations is not.

Every COO builds their own

Every Head of Digital and CMO we've interviewed for the Plaith Operational Benchmark has built a version of this themselves. Custom prompts. Bespoke n8n flows. Glue work that doesn't compound, because the moment the agency or freelancer leaves, the work leaves with them.

The big consultancies have noticed the gap, but they're servicing only the top of it. Deloitte launched its Centre of Excellence for Oracle AI Agents in October. Accenture, KPMG, and McKinsey have all built large GCC AI advisory practices in the past eighteen months. Their work is aimed at sovereign-scale and large-enterprise clients, where multi-year transformation programmes can absorb the cost of a custom build.

For the 94% of UAE businesses that are SMEs, and for the mid-market consumer brands building inside Saudi Arabia's $25 billion e-commerce market, that procurement model doesn't fit. A toned-down version of it doesn't either.

What changes in the next two years

Operational AI is different from previous enterprise technology in one important respect. Its defining property is ultra-personalisation at any scale. A small consumer brand running across three markets in four languages can have the same operational depth as a sovereign client, if the infrastructure is built for that shape.

SMEs and mid-market businesses don't need a smaller version of enterprise AI. They need infrastructure built for the way they already operate. That's a different shape entirely from what the big consultancies are building, and a different shape from what the licensed-by-the-seat AI tooling industry is selling.

This is the direction the relationship between AI and consumer brand SMEs should be moving in. It's the direction Plaith is pioneering.

The question worth sitting with is which parts of an operation should be running on infrastructure nobody has to procure or rebuild themselves.

Have a question about this?

We're happy to talk through the specifics. No pitch, no agenda.