Sleep App
Combines sleep and fatigue data into one picture and suggests decisions about training and recovery.
Automation and AI for repeatable work in your company
Let's build one system that connects your data, rules and tools. Your team knows what happens next. And you only make the key decisions.
Decisions about money, risk and customers always stay on your side.
Sound familiar?
My manifesto
Adding automation and AI to a broken process only speeds up the mess. First I fix the data, rules and exceptions, then I build the system.
Case study · LetterOperators
We did not start with AI agents. At first more than a dozen copywriters wrote the newsletters. Then we moved through custom GPTs and an automated workflow. In the end we built a complete operations platform with Claude based agents.
The platform gathered news from managed sources, scored relevance for each client, wrote in the client's voice and prepared the send for Beehiiv. A human read the final draft and clicked approve.
That is what a company that stopped waiting for its owners looked like. We built a cloud platform the new owner's team can run from anywhere. In February 2026 we sold the company together with the system.
One context · many subsystems
Each module uses the same company context and leaves a usable history.
Data, definitions, rules, history and exceptions become a shared operating base.
Email, sheets, CRM, documents and apps hand over the data the next step needs.
The system prepares the context. A person decides where money, risk or a customer is involved.
After approval the work moves on. The system records what it did and why.
Inputs
Shared company context
Data, definitions, rules, history and exceptions in one place.
Agents and subsystems
Human gate
You approve when money, risk or a customer is involved.
Outcome
The work moves on. The system records what it did and why.
Other builds
Combines sleep and fatigue data into one picture and suggests decisions about training and recovery.
Brings trip data, costs and margin into one place and prepares the offer for the team and the customer.
Analyses a TikTok Ads account and proposes changes that help lift the channel's ROI. It learns from past results.
An operations hub for brands selling on Amazon. Listings, competitor research and market research run in one system that works in the background.
Conversation · Workshop · Build
We talk about your business problems. We check where the company loses the most time and whether you need a partner to build the system with you.
Book a 30-minute conversationWe map the operation, decisions and exceptions. You get a report with the recommended first stage. I invite companies to the workshop only when I see a real path to a working implementation.
A written scope, test cases, rollout, training and documentation. The system stays with you.
A first system usually costs from 4900 to 15000 PLN. You get the exact scope and price in the report after the workshop.
I run at most two active implementations at a time. That is why I take the first conversation seriously on both sides.
Book the first conversationMaciej Marek · operator and builder
For six years I worked in finance and procurement at Akamai, GEP and HP. I reconciled invoices, ran supplier contracts and system rollouts. I saw from the inside where companies really lose time and money.
Then I built marketing automation for online stores. The systems took over campaigns and sends that teams had been doing by hand.
With Richard I co-built LetterOperators. I ran the operation and the system. We went from more than a dozen copywriters to a platform run by two people. In February 2026 we sold the company.
Today I do the same for other companies. I straighten out the operation, build the system, train the team and hand over the documentation. Decisions about money, risk and customers stay on your side.
Before the conversation
The conversation and the workshop are free. A first system usually costs from 4900 to 15000 PLN and takes two to six weeks. You get the exact quote in the report after the workshop.
Before the build we agree which data the system really needs, where it may be processed and who has access. I document the integrations and control points.
No. After the workshop you get a report and a recommendation. Only then do both sides decide whether we build together.
The agreed system, documentation, acceptance criteria, training and a recording of the handover. The proposal defines accounts, access, licences and any further support.
We write down test cases, correction rounds and consequences before the build. If the criteria are missed, we document the cause and decide together whether we fix, change the scope or stop.
Before the build we agree which data the system uses, which decisions it supports and where human approval is required.
A 30-minute conversation
Instead of an agency and yet another tool you get a partner who knows operations from the inside and takes responsibility for a working system. If both sides see the point, we book the workshop and you get a report with a plan for your first system.
Book a 30-minute conversation