AI Product Consulting · Islamabad → Worldwide

I cannot writeproduction code.I have shipped six products.


Not a paradox. Building software became a specification problem, and specification is an operator skill.

Or get the weekly newsletter

Demo only: no backend is connected yet, nothing is stored or sent.

[PLACEHOLDER: real subscriber count] operators and founders already subscribed

Weekly. One click to unsubscribe.

  • NIC Islamabad · Cohort 12
  • Co-founder & CEO, Triisum
  • 6 products shipped
  • $1M seed raise run

Watch first

The twelve-minute version of this page.

Everything on this page, explained out loud, with the screen share. · 12:04

[PLACEHOLDER: unlisted YouTube video ID, embedded via youtube-nocookie.com]
Book a free 30-min strategy call

The 12-minute version

Everything on this page, explained out loud, with the screen share. · 12:04

No video connected yet.

Add the unlisted YouTube video ID to vsl.embedUrl in src/data/landing.ts. It will embed through youtube-nocookie.com so the page stays cookie-free.

No pitch on the first call. If I am not the right fit I will say so in the first ten minutes.

Nothing loads from YouTube until you press play. No cookies are set before that.

Who this is for

Three people get disproportionate value from this. You are probably one of them.

If you are not on this list, the newsletter is free and there is no reason to book a call.

How I found this

For four years I was the guy who could not build his own product.

I co-founded Triisum in 2021 and ran it as CEO. I did the parts an operator does: 50+ customer interviews, the white-label partnerships, the OTA integrations, the NIC Islamabad cohort, an eight-person team across engineering, QA and ops. What I could not do was build any of it. Every idea I had went into a queue, and the queue was other people, and the queue was always full.

You learn a specific kind of patience running a company you cannot personally build. You get very good at writing specs, because the spec is the only lever you have. You also get very good at watching a two-week estimate become six, because the thing in your head and the thing in the engineer's head were never the same thing and neither of you found out until the demo.

So when the AI coding tools arrived I did what every non-technical founder does. I typed what I wanted in a sentence, the way you would ask a person. And I got back what everyone gets back: something that runs, looks roughly right, and is wrong in a way that takes longer to unpick than to rewrite. I concluded the tools were for engineers and went back to the queue.

What changed was realising I had been treating it like a search engine (short question, magic answer) when it is closer to an extremely fast, completely literal contractor with no memory of yesterday and no ability to ask you a clarifying question. So I stopped writing prompts and started writing **the brief I had already spent four years learning to write for humans.** Constraints. Edge cases. The file it goes in. What "done" means. What it must not touch.

The output stopped being wrong. Same model that afternoon as that morning: the difference was that for the first time the specification was actually complete, and I only knew that because something had executed it literally enough to expose every gap. Since then I have shipped LittleOS, Liquid Finance, Safar, ThumbSnap AI, an autonomous publishing pipeline, and the Eco-Ambassadors app. I still cannot write production code.

That is the whole argument. The four years of operator work were not the obstacle to building; they were the qualification for it, and I had it backwards the entire time. It took me months of doing this badly to see that. It takes my clients about three weeks, because they get to skip the months.

The thing I thought disqualified me was the thing that qualified me.

The 12-minute version

Everything on this page, explained out loud, with the screen share. · 12:04

No video connected yet.

Add the unlisted YouTube video ID to vsl.embedUrl in src/data/landing.ts. It will embed through youtube-nocookie.com so the page stays cookie-free.

The one thing

If building software is now a specification problem rather than an engineering one, and if specification is a skill you already built by running a business, then being non-technical stopped being the reason you cannot ship, and quietly became the reason you can.

That is the only claim I am making and the only one that has to be true. If you believe it, the question stops being whether to build this capability and becomes how quickly you can install it. If you do not believe it, no offer, bonus or guarantee below is going to change your mind, and the honest recommendation is to read the newsletter for a few months and decide for yourself. The six products in the work section are the whole of my evidence. Go and look at them before you decide.

Consulting

Three ways to install this. They are the same system at three depths.

Every engagement starts with the same free call, and I will tell you on that call which of these you actually need, including when the answer is none of them.

The risk is mine, not yours

If you finish the first week of a Build Sprint and do not believe it will pay for itself, tell me and I refund it in full. I have never had to, but the offer is real and it is in the contract: I would rather lose a fee than have someone tell five people it was not worth it.

Courses

Or learn it yourself, at a tenth of the price and four times the pace.

Consulting does not scale and most people do not need it. The courses are the same system, self-serve. Start with the free one: if it does not change how you work in the first hour, do not buy the others.

Course titles below are proposals. Rename anything.

Proof

Five clients. Here is what they say.

These slots are empty on purpose. Real quotes, with permission, or nothing (see CONTENT.md).

Products shipped
6
Clients advised
5
People trained on AI tools
15+
Following on LinkedIn
9K+

Work

Six products. None of them written by me.

This is the section that has to carry the whole argument, so it is the one with the least spin in it. Every project below is real, most are public, and the stacks are listed because a sceptical engineer reading this page will want to check.

Tools

The stack I actually pay for.

Short list on purpose. I review things I use on real work, and I say when something is not worth it: a directory where everything scores well is an advert, not a review.

Some links below are affiliate links, marked as such. If you buy through one I may earn a commission at no extra cost to you. It does not affect the rating: I have left tools I was offered money to include off this list entirely.

Watch

The whole argument, out loud.

If you would rather watch than read, start here. Screen share included, so you see the actual specifications rather than slides about them.

Nothing loads from YouTube until you press play. No cookies are set before that.

The newsletter

What is actually working in AI this week.

One email. What shipped, what is hype, and the one thing worth your Saturday morning. Across AI, coding, writing, images and Web3. Written by me, not generated and sent unread.

Weekly · unsubscribe in one click

Demo only: no backend is connected yet, nothing is stored or sent.

No spam, no list sales, no third-party sharing. Stored in the EU.

Writing

Longer arguments, worked through properly.

Where the newsletter is what happened this week, these are the pieces I keep sending people.

All writing

Before you book

The objections, answered honestly.

I tried this. It wrote code I had to throw away.
Yes. That is the normal first experience and it is the reason most people stop. What you learned from that is that a one-line instruction produces one-line-quality work (which was always true of humans too, we just never noticed because humans quietly fill the gaps and then resent you for it). The fix is not a better model or a longer prompt. It is a structured brief with constraints and a definition of done. That is the thing I teach, and it is most of the first week of any engagement.
I am not technical. Is this actually for me?
It is the only group I am genuinely the right person for. I am not technical either: I describe myself on GitHub as a non-technical person teaching other non-technical people, and the six products in the work section were all shipped without me writing production code. What you do need is the ability to describe what you want precisely, which is an operator skill and one you probably already have if you have ever written a decent brief. The one thing I will not tell you is that you can ship a production codebase with zero technical judgement. You cannot, nobody can, and anyone selling you that is selling you something.
My engineers will hate this.
Some will, and usually for a good reason: they have watched a manager use AI output as an argument for a deadline that was never realistic. The adoption work is mostly about not doing that. In practice the engineers who resist hardest at the start become the strongest users, because the system rewards exactly the thing they have been asking for since they got there: actual specifications instead of a Slack message and a shrug.
I do not have time to learn a new system right now.
Then start with the free call and the free course, and spend nothing but forty minutes. If forty minutes is genuinely not available then the honest answer is that this is not your bottleneck this quarter, and I would rather tell you that than take a fee.
Why you? There are a hundred AI consultants now.
There are, and most of them are either engineers who never had to run a business or operators who have never shipped anything. I co-founded Triisum and ran it as CEO for four years, ran a $1M seed process at Propellus with a 30+ investor pipeline, trained 15+ people on these tools inside a real company, and shipped six products of my own while remaining non-technical. I am also in Islamabad rather than San Francisco, which mostly means my rates are not padded by a Bay Area lease. The work section lists the stacks: go and check them, then judge.
What if it does not work for us?
Then I refund the sprint: first week, no argument, in the contract. I would rather lose a fee than have someone tell five people it was not worth it.

Last thing

The call is free and I am not going to pitch you.

Thirty minutes. You describe what you are trying to ship and where it is stuck, I tell you what I would do about it. If a sprint is the right answer I will say so. If it is not, I will say that too: most of these calls end with somebody getting a useful answer and not spending any money, and that is a completely fine outcome for me.

P.S. If you only take one thing from this page, take this: the next time you get bad output, do not rewrite the prompt. Write the brief you would have handed a contractor who cannot ask you questions. The difference will be obvious inside an hour, and you will not need to hire anybody to see it.