Operations systems
Process mapping so the whole flow fits on one page, workflow systems across the tools you already pay for, and reporting that assembles itself. The spreadsheet three people update by hand, retired.
Operations studio, Lisbon
Most companies lose whole working weeks to work that software should be doing. We find those hours, then build the systems that give them back: operations, revenue, marketing, e-commerce, finance and legal, and payment rails that move real money across 31 chains.
What we do
We work across the parts nobody puts on a pitch deck. AI does the heavy lifting where it genuinely removes work, and stays out of the way where it does not.
Process mapping so the whole flow fits on one page, workflow systems across the tools you already pay for, and reporting that assembles itself. The spreadsheet three people update by hand, retired.
Assistants that know your documents, your processes and your tone. Intake that reads, sorts and routes the work before anyone opens an inbox. Applied to the work, never bolted on top of it.
Pipeline structure and data you can trust, scoring so attention goes where it pays, and follow-up that happens whether or not anyone remembers. Less leaks out between interest and reply.
Positioning that survives contact with a real buyer, campaign planning and execution, content that compounds rather than evaporates, and measurement so spend is a decision and not a habit.
Technical SEO, site structure and page speed. Product and category pages built to rank and to sell. Checkout and conversion friction, found and removed. Demand that arrives without a media invoice.
Reporting packs that build from source data, reconciliation and close routines, contract and obligation tracking, and compliance records ready before anyone asks. A team we have worked with long enough to treat as one.
Screening and scheduling that moves in hours, structured interviews so decisions are comparable, onboarding that makes someone useful in days, and contractor operations at scale. We run a 43-person bench day to day.
The sharp end of the studio, and the piece we can prove hardest. Payout rails, on-chain reconciliation and double-entry ledgers for teams that move real money. It is running in production today, across 31 chains.
The detail is belowPayments, the sharp end
Crypto payroll, marketplace and creator platforms usually run fiat payouts at 100% and token payouts somewhere in the nineties. The gap is where money, support hours and trust leak out. It is rarely one bug. It is the seams between systems that each work on their own. We build the four layers that close them.
Multi-chain dispatch that stays idempotent under retried webhooks and duplicated jobs. Custody-provider agnostic, so a Safe, an exchange and an MPC wallet all look the same to the ledger.
Each transaction verified against what actually landed, per token, per chain, independent of the provider that claims it was sent. The record and the chain agree, or a human hears about it the same hour.
Double-entry, with balances derived rather than stored, and partial-settlement rules written down and enforced in code. Finance closes the month from the ledger, not from a spreadsheet next to it.
Aging digests, stuck-payment detection, and monitoring that pages someone before a customer notices. The part most teams never get to, and the part that makes the other three trustworthy.
Start free. Send a link to your docs or status page and we reply with two pages: where your payouts leak and the three changes that stop it. From there, a fixed-fee audit, a fixed-scope build, or a monthly retainer where we own the reliability of your payouts.
Proof, not slides
Everything on this page rests on work that is live. The most measurable piece is one production system, written end to end: payout dispatch, on-chain reconciliation and the ledger behind them, plus the migration off Google Apps Script onto Postgres and Node that made it possible.
Before that. Operations at serious scale in digital assets: the first AML and KYB framework at Starknet Foundation, built with Sumsub. A treasury above $100M on Safe multisigs. Grants for 500+ teams, run end to end. The habits came from an environment where a missed record is a regulatory problem, and we bring the same standard to a shipping process or a reporting pack.
How we work
AI does not fix a business. Understanding how the work really happens does. The tools come after that, and they matter less than most people selling them suggest.
Not only the leadership. The person who rebuilds the same report every Monday usually knows exactly what is broken, and has rarely been asked.
The reporting complaint is often a data collection problem. The hiring delay is often a scheduling problem. We keep going until the cause stops moving.
You know what you are getting and what it costs before we start. No open-ended retainer that bills for months before you see anything.
Something working in front of your team beats a perfect plan. We adjust from what they actually do with it, not from what we assumed.
Documented, explained, and owned by you. If you still need us a year later, we would rather it was for something new.
How we judge it. Six months after we finish, your team is still using it and has changed it themselves without calling us. That is the result we are aiming at.
Who this is for
Not a fit. Anyone who wants a tool before they want to understand the work, institutions whose procurement runs longer than the project, and anyone who wants regulatory advice. We build the systems your compliance team relies on. We do not sign off on your policy.
Who you work with
You talk to the people doing the work. That is the whole structure, and it is deliberate.
Small on purpose. It is the reason we can say no to work that is not a fit.
Start here
A short conversation, no pitch. If we are not the right people for it, we will say so and point you somewhere better. Useful things to bring: what the process is, roughly how many hours a week it eats, and who it frustrates most.
kisubimalcolm@gmail.com. Lisbon, working with teams across Europe and the US.