Company
Didrah is built for repeating distribution-sales decisions
An agentic operating layer, not a content studio and not a dashboard company. Our audience is the sales and commercial manager in distribution.
Who it is for
Someone who has to know this morning which brand, campaign, or region is the priority, and cannot wait for the monthly pack. If your problem is brand awareness on social, this page is not for you.
Two tracks. One shared problem.
One of us has spent a decade on the receiving end of a report that came too late. The other has spent a decade building the systems that make “too late” unnecessary.
And from exactly those two angles we reached a shared observation: in distribution companies, data is not scarce. What is scarce is data arriving at the moment of decision. Weekly repeating decisions are still made in Excel, on experience.
Stable roles at Golrang and in AI/ML would have kept us solving slices of this. We are building it as a startup because the full loop (judgment on the commercial floor, running in-country, unlocked from a foreign stack) only gets built as its own company.
What we got wrong
We thought the hard part would be the models. It wasn't. The hard part was that real distribution data is scattered and inconsistent, and building something that actually works on it took longer than writing a clever algorithm.
Team
محمد شکرزاد
CEO
For ten years I was the manager waiting on the report: deciding which brand or region gets attention with three slides and no time to ask a fourth question. That decade spans category management across dozens of stores, building a brand from zero, and leading trade marketing with hundreds of field representatives and a multi-million-dollar annual budget. Today I own brand strategy across more than five consumer brands at Golrang Industrial Group; exactly where a commercial decision becomes execution at the point of sale.
رضا شکرزاد
CTO
Every system I've shipped ends the same way: turning something ambiguous into a decision a machine can make reliably, and reliably enough to trust it. Two master's degrees: data science from Radboud in the Netherlands and industrial engineering from Sharif; plus deploying machine-learning models at Henkel Amsterdam on an FMCG production line; founder of an education platform with 400,000 learners. That specific background is what makes in-country, model-agnostic deployment possible rather than aspirational.
Let’s make your processes agentic
Where do we start?