Three of the thirteen ventures on this site did not work. Mokaab never launched. WritingChex wound down. CompanyHouse closed. They sit in the same grid as the ones that shipped, with the same dates and the same status labels.
People assume this is modesty. It is not. It is a measurement problem. If you only show me your wins, I have no idea what your wins cost or what you were willing to abandon. A portfolio without failures is a portfolio you cannot calibrate.
What actually broke
The three broke in three different places, and the differences matter more than the fact of failure.
CompanyHouse failed at extraction. The idea was straightforward: registering a company in Iran requires an expert because the process is deliberately opaque. The expertise is the product and the bottleneck. So productise the expert. I modelled the specialists' tacit knowledge into a guided, question-based workflow that generated the filings, and worked with legal experts to convert rules, decision points and exceptions into product logic. It ran for a year and closed in 2016.
The failure was not the software. It was that the knowledge did not live in the form I needed it in. Ask a specialist how they do it and they describe the happy path. The value is in the exceptions — the cases where the rule bends, where a filing gets rejected for a reason nobody wrote down. Those exceptions are invisible to the person who holds them, because to them they are not exceptions.
That is the lesson: productising expert knowledge is a knowledge-extraction problem long before it is a software problem. Until you can model the questions a specialist asks and the exceptions they know, you do not have a product — you have a service with a login screen. It cost me a company to learn that.
Mokaab failed at priority. I co-founded it as a gold and silver investment platform and architected the full product framework — buy, sell, hold, buyback, pricing, invoicing, delivery and customer account journeys — with the buyback obligation designed in from day one rather than bolted on. It never launched. The parent group's priorities moved to physical operations before the platform shipped.
I want to be precise about this one, because it is the failure most tempting to excuse. The thesis was that in Iranian gold the trust gap, not the technology, is the market. That thesis was not what killed it. It died because the organisation funding it needed something else more urgently, and I was building a digital product inside a group whose weight sat in physical trading. It is still a failure and it is still partly mine.
WritingChex failed after being right. An IELTS candidate outside a major city has no examiner, no feedback, and no way to know why their writing scores what it scores. We built an AI feedback engine that simulated the exam, analysed the response and returned examiner-shaped feedback, then designed the improvement loop around it. The MVP was validated on real submissions and real users, and we got the first paying customers — the first evidence anyone would pay for AI-generated feedback. It wound down in 2025.
That is the uncomfortable shape. Validation is not survival. We proved demand and did not convert that proof into a business. I will say the one thing I am mildly proud of inside a failure: I was honest with the team about where the model could not be trusted.
What did not die
Mokaab's architecture is still running. The architecture and the market model survived and were folded into Shalize's operating system. Shalize was a physical gold and silver business running on WhatsApp messages, memory and trust, with volume climbing faster than the process could hold it. I rebuilt the operating system end to end — enquiry, order, invoicing, payment verification, fulfilment, follow-up and buyback — and put a CRM under it. Each of those steps got one owner, one definition and one place to live. Daily volume moved from 0.5 to 10 kg per day, and it recorded 4,000+ verified purchasing customers.
The buyback obligation designed into Mokaab from day one is the same commitment written into Shalize's product rather than its sales pitch. The unlaunched product supplied the design of the launched one. I did not know that at the time.
This is the honest reason to keep failures visible, and it is not sentimental: they are load-bearing. CompanyHouse is why I ask knowledge-extraction questions first. Mokaab is why Shalize's operating system did not have to be invented from nothing. WritingChex is why I do not let a model's confidence stand in for its accuracy.
The narrower version of the same discipline
Saying what did not work also means saying what you did not do. NoJahan launched and validated its early model, then scaled without me. So the site says exactly that: I am only claiming the zero-to-launch work, not its long-term trajectory. It would be easy to let a reader assume otherwise. Nobody would check.
The same applies to things still in flight. Sarzamin-e Emkan and Invest Iran are described as in active development, because that is what they are. Ketabno launched in roughly three months and partnered with the Ministry of Education and the national Shad platform — stated because they happened, with nothing further claimed about what they produced.
And it applies to how I describe the years. Aghigh is the longest chapter, and it is where I learned company building, governance and operational scaling — by doing all three badly first. That sentence is on my own experience page. I put it there because the alternative version, the one where a long stretch of learning reads as a long stretch of competence, is a lie that gets found out in the first serious conversation.
Every figure on this site says where it came from and whether it was audited. Most of them were not. The cost of saying what did not work is a paragraph of discomfort. The cost of not saying it is that nothing else you say can be checked.