Blog

Material Intelligence β€” stop ordering twice, stop running short

AI-driven material planning from past projects: how to flag over-ordering and stock-outs before they bleed margin.

Eng. Amr Shoieb6 min read
materialsaicost-savingsuae-construction

Walk a Gulf construction site at any month-end and you'll find two patterns at once: a yard full of materials nobody's ready to use, and a foreman on the phone trying to chase the wrong material that's suddenly short. Both eat margin. Both come from the same root cause: nobody has the full picture in real time.

Where material planning actually fails

The textbook answer is "plan to schedule." In practice, three things break that plan:

  1. The schedule changes weekly. Concrete pours slip; finishing sequences flip. The procurement plan that was right on Monday is wrong by Thursday.
  2. The take-off is optimistic. Most BOQs under-call on consumables β€” joint compound, sealants, dowels, fixings. A trade lead orders "the BOQ quantity plus 5%" and runs short anyway.
  3. No tenant-level memory. Your last 5 mosque projects all over- ordered the same SKU of tile adhesive by 12%. Nobody on your team remembers that across projects.

Material Intelligence is the layer that fixes problem 3, which makes problems 1 and 2 manageable.

What "AI material planning" actually means

It isn't a chatbot. It's:

  • Consumption modelling per tenant. ORKSTRA tracks the gap between ordered and consumed quantities across every project, per SKU, per trade.
  • Variance prediction. When you raise a new PO for the same SKU on a similar project, the engine surfaces last quarter's variance. "Last 5 mosque projects: ordered 12% more copper than installed. Adjust this PO?"
  • Theft + leakage signals. Variance that's outside the consumption distribution gets flagged separately β€” not as "over- order," but as "unaccounted." The pattern catches yard pilferage early.
  • Stock-out alerts. When material draw-down rate exceeds plan, the engine warns before the site runs out β€” not after.

What this is worth

On Aletlala's UAE portfolio (which we used to calibrate the engine), Material Intelligence cut material wastage by roughly 12% across the finishing trades, where it matters most. On a typical AED 80M contract with 35% material content, that's AED 3.3M of recovered margin.

If your stock yard looks like the picture above on month-end, that's not a procurement-discipline problem. That's a missing-system problem. And it's exactly the gap ORKSTRA's Material Intelligence was built to close.

Ready to streamline your projects?

ORKSTRA is launching soon. Join the waitlist for early access β€” first 100 customers get 30% off.