In July 2011 my friend Jetmir and I started building what became Sindikata e Algoritmeve te Pavarura (in English, the Syndicate of Independent Algorithms), registered as S.A.P. LLC, with the company Trembelat joining as a third stakeholder. We formally established it in Prishtina in December 2012. I served as managing director from the start until March 2013, when personal life events took me to California.
Back then the deal was simple. A client came with a need: a website, an online store, a new brand. We wrote a scope, agreed on a price and a deadline, built it and launched it. Then we shook hands and moved on to the next one.
At the time we called that success. Looking back from the other side of the world, I see it differently: a launch was where our work ended and where the client’s real problem began.
The project was never the product
I can count on one hand the clients who needed “a website.” What they needed was more customers, more online sales or a brand people trusted. The website was our best guess at how to get there.
A guess needs testing. Once a site is live, the questions start. Which page loses people? Why do visitors add to cart and then leave? Which message actually brings in leads? The project model answers none of these. It ends at launch day.
So the site sat there. Months later the client came back with a new brief, and often it was a redesign. We started from scratch, with a new quote and a new launch. Neither of us learned much along the way.
What clients actually needed was a loop
The companies that grew worked in a loop instead of in projects:
Ship → Measure → Learn → Adjust → Ship again
Small changes every week beat a big rebuild every couple of years. But for most of our clients in Kosovo and the region, a loop was too expensive to run. It took a designer, a developer, a marketer and an analyst, all paying attention all the time. A small business couldn’t hire that team, and a monthly agency retainer to cover it felt like a luxury.
So the loop had no owner. We left after launch, and the client was too busy running the business to keep it going.
AI changed the economics
In the years since I stepped away, and especially in the last few, the cost of execution has collapsed. Writing a landing page variant, building an email flow, analysing a month of analytics, producing product descriptions for 500 items: work that used to take our team a week now takes an afternoon. Some of it is now something a founder can do alone at night.
Watching this from California, I keep thinking about how many of our early client problems would look different today.
Some people see that as the end of agencies. I see it as the end of a certain kind of agency: the one that sells hours and deliverables.
When execution is cheap, value moves to:
- Judgement: knowing what is worth testing and what is noise.
- Context: knowing the market, the customers and why things didn’t work last time.
- Continuity: someone who stays in the loop and doesn’t disappear after launch.
So who owns the loop now?
There are three honest answers:
- The founder owns it. With AI tools, a determined founder can run much of the loop alone. Sometimes the best thing an agency can do is set up the systems, train the team and step back.
- An in-house team owns it. One generalist using AI now covers what three specialists used to.
- A partner owns it with them. Instead of a scope and a deadline, client and partner agree on a goal (sales, leads, conversion) and work on it every month. The partner uses AI to move faster and spends its time on decisions, not production.
What no longer works is selling a launch and leaving.
What I would do differently
If I were starting Sindikata today:
- Fewer one-off projects, more ongoing partnerships with clear goals and a monthly rhythm.
- AI in every part of the process, so clients pay for thinking, not typing.
- Reporting that answers “what did we learn?” instead of “what did we deliver?”
- Honesty about when a client doesn’t need us and can run the loop themselves.
The last point may sound strange coming from an agency founder, but I believe it’s the only position that holds up long term. If a client can run the loop alone, they will find out sooner or later. Better to help them get there and stay as the partner they call for the hard problems.
Closing reflection
In the early days of Sindikata we measured ourselves by what we launched. The agencies that win from here will measure themselves by what changed for the client after launch.
Execution is no longer what’s scarce. Momentum is, and AI has made it possible for far more businesses to keep it going, whether they do it themselves or with a partner.
I’m proud of what Jetmir and I started in Prishtina, and grateful to everyone who carried it forward.
Write a comment