Everyone told me to pick a lane.I picked them all.
Berke Tok · Technology Consultant, Dubai
01Product management
Deciding what will not be built, not what will.
02Digital strategy
A strategy that produces no decision is not a strategy.
03Banking applications
Building something usable inside the regulation.
04Exchange & crypto infrastructure
Thousands of orders a second, zero tolerance.
05Artificial intelligence
Not the demo. The thing that survives production.
06Brand
The sentence people use when they describe your product.
Six disciplines.One system.
Most problems sit between two of these. That is why they are one practice and not six services.
01Roadmap · Discovery
Product management
A roadmap is not a wish list. Most of the job is turning down good ideas and being able to say why.
02Positioning · Scope
Digital strategy
The output is not a deck. It is who does what by when, and what was given up to make that possible.
03Compliance · KYC
Banking applications
Compliance requirements do not have to kill the product. Usually what is bad is not the regulation but the lazy answer given to it.
04Matching engine · Custody
Exchange & crypto infrastructure
Matching engine, wallet infrastructure, liquidity. There is no such category here as “works most of the time”.
05Evals · Retrieval
Artificial intelligence
An impressive demo takes a week. Working out what happens when the model is confidently wrong is the actual job.
06Positioning · Voice
Brand
Brand is not a logo. It is the one line a customer uses to explain you to a friend. If you do not write it, someone else will.
07 · BETWEEN TWO
Not sure which one you need?
Neither are most clients. Describe the problem; the discipline follows from it.
Start a project
Three moves.The order does not change.
Most briefs arrive as a proposed solution: “we need a mobile app.” The first job is finding the real question underneath. Half of all projects change shape here and some of them end here, which is also a good outcome.
Not a deck. A working piece. If there is nothing you can open and try within two weeks, the problem is the plan, not the team.
Good consulting makes itself unnecessary. If it stops working once I leave, the work was not finished.
Systems that runafter the meeting ends.
03 selected cases
01 · 2026
AlmaraClaims settled before the file gets cold
- Sector
- Health insurance
- To production
- 18 weeks
02 · 2026
AssemblyContributor payouts with a paper trail
- Sector
- Distributed organisations
- To production
- 13 weeks
03 · 2026
AttestVerify once, reuse everywhere
- Sector
- Digital identity
- To production
- 19 weeks
Interfaces and identitiesacross all six disciplines.
12 projects
HarazDigital strategy · Hospitality
OstraExchange & crypto infrastructure · Digital assets
CadenceProduct management · Product tooling
BriefArtificial intelligence · Legal
KoraBanking applications · Retail banking
SableBrand · Fragrance
WaylaDigital strategy · Mobility
The long version,kept short.
The fastest way to understand how something works was to break it and put it back together. What I took from that period was less technical knowledge than a habit: not talking about a system before understanding it.
I watched technically flawless work go unused. Nobody had asked the right question and nobody had written the sentence that explained the product. Then I realised the gap was the actual job.
Consulting alone only goes so far. The agency exists to close the distance between “you should do this” and “we did this”.
Crypto exchanges, banking applications, health technology, AI products. What they share is not the sector but the shape of the problem: something standing between two sides that do not understand each other.
Fixed opinions
04- Complexity is not a sign of achievement. It is usually a form of postponement.
- A good product is hundreds of decisions the user never notices.
- If you cannot explain it, you probably do not understand it yet.
- Fast and rushed are not the same thing.
- +971 52 577 2724
- mail@berketok.com
- in/berke-tok
- GitHub
- berketok-git