Pirxey Time
An internal operating system for a software house: time tracking, projects, client billing and an HR module. The build cost about PLN 50k, while comparable SaaS would cost about PLN 162k a year.


A CAIO is the person in a company who makes AI deliver: picks the process, calculates the return and takes the rollout to production. With me, you get that without adding a seat on the board. One person accountable for the result.
Each project in brief: the problem, the solution, where AI does the work and how long it took. Some also have a full case study with the numbers and with what broke along the way.
An internal operating system for a software house: time tracking, projects, client billing and an HR module. The build cost about PLN 50k, while comparable SaaS would cost about PLN 162k a year.


A system where AI agents take a ticket from plan through code and tests to code review. A human makes two decisions per ticket, and about 50 tickets a week go through the factory.



Proprietary machine learning and computer vision technology that reads emotions from faces in a video stream from an ordinary camera. It let companies measure viewer and customer reactions as they happened, while people watched an ad or talked at a counter.

A family app that turns a child’s daily routines into a game with tasks, points and rewards. Children come back to their daily plan more willingly, and parents set tasks and rewards in one place.






An AI agent for a tender team that moves data from client spreadsheets with different layouts into one bid format. The Power Platform experiment showed that further changes were faster to make in code with agents than in low-code.

A personal task system with an Inbox, priorities, daily triage and a task draw, which AI agents can use through MCP. I plan my day in it with agents and change the app when the way I work changes.
AI rollouts and automation
A family app for sharing how you feel: a quick check-in, a private journal and an evening ritual with a tarot card and an AI reading. It helps us know how our family feels, even on days with no time for a longer talk.
Digital products and integrations
A food-ordering PWA for a food truck chain, with online payments and a loyalty programme. Orders from the app, the till and the phone all land in one POS system.
Digital products and integrations
A POS system for food businesses running a Central Kitchen model. The till, the app and the voice agent sell from one menu and record orders in one database.
Software built for a business need
A personal finance system: spending control, planned purchases and Pay Yourself First savings, which AI agents can use through an API and MCP. An agent analyses the budget and makes changes within the permissions it has been given.
Digital products and integrations
A professional medical records system for aesthetic cosmetology clinics. Patient records, treatment photos and visit history live in one place, with data protection that meets GDPR.
Software built for a business need
A system for managing work at a shipyard: job packs, working time reports, employee reviews and bonuses. AI agents made further development of this legacy system economically viable again.
Software built for a business need
A system for banks and service points that turns advisor and customer smiles into staff rewards and support for charities. In the PKO BP pilot, advisors smiled 2.3 times more often.
AI rollouts and automation
A browser-based tool that measured viewer reactions to ads, trailers and casting candidates through facial expression analysis. It showed which scene and which actor engaged viewers before a company spent its campaign budget.
AI rollouts and automation
A mobile mood diary for iOS and Android: a short mood test, situation tags and a mood history chart. It helped people see what affected their mood and share that history with a therapist or someone close.
Digital products and integrationsI start with the process and the money and choose the technology last. I work out what manual work costs today and show where AI will pay back within months. Where it won’t pay back, I say so plainly and suggest a simpler solution.
No packages. We agree the scope after a call, and each service lists the projects that back it up.
I map the process, work out the cost of manual work and roll out agents and automations where they pay back fastest. From pilot to production, with a human in the loop where it matters.
Internal systems that replace spreadsheets and several subscriptions at once. I design them around your process and the people who use them. You own the code and the rights.
MVPs for founders and products ready for users: architecture, payments, integrations, PWA. From prototype to a version you can sell.
You pay for actual working time. Staged payments, no hidden costs.
A review of your idea, architecture or AI rollout plan before you spend the budget.
From about fifteen to forty working days, from the first call to production.
A 30-minute call, then a process map. We work out what manual work costs today and where AI will really pay back.
A clickable prototype on your data. You see the whole flow before any production code gets written.
Production architecture, tests and integrations. You see progress every day on a working deployment.
Go-live, monitoring and further development. You own 100% of the code and the rights.
speakerKoszalin 2014 and Wrocław 2019. On technology, kindness and what machines teach us about people.

Hover to see the photoI have built companies and written their code. I started at an interactive agency, then came the Quantum CX startup, with rollouts at PKO BP, Orange and Santander, and the Seneka Fund. That is why I talk to the board in terms of results and to the team in terms of architecture.
I write and speak about AI in organisations. I take invitations to talks, podcasts and media by email.
Usually somewhere between fifteen and forty working days. After the first call you get a plan with stages and a day estimate.
PLN 2,500 per working day, and consultations at PLN 800 per hour. You pay for actual time, in staged payments, with no packages or hidden costs.
Today’s models are very strong, and they need supervision. I take on the architecture, the review and the responsibility for production. 20 years of programming means I know what AI will miss.
You do. No licences, royalties or vendor lock-in. You can keep developing it with any team.
Security by design, RLS, dependency audits, e2e tests. I have worked for banks and a hospital, so I know what compliance means.
For large, long projects I work with Pirxey: 70+ engineers, where I lead AI development.
30 minutes is enough to judge whether and where AI will pay back in your company.
Pick a time for a call