Someone finally runs AI in your company

Your External AI Director

Together with your board we pick the initiatives that make sense, drive them to working solutions and measure the effect. You get one person accountable for the whole area, and a team that delivers rather than only recommends.

No sales deck. We talk about one area that is stuck today and tell you plainly whether we can help.

Michał Stanisławek and Rafał Garbacz, founders of dyrektor.ai
Michał Stanisławek / Rafał Garbacz, founders of dyrektor.ai

When you need one

AI is not stuck for lack of tools. It is stuck because decisions and follow-through are missing.

Every department has a different idea.

Teams test different tools, but the company has no shared criteria and no single list of priorities.

Tests never turn into real work.

Business, IT and vendors wait for each other because nobody owns agreeing and doing the next step.

The board has no basis for a decision.

Costs, risks and expected gains are judged separately. It is hard to decide what to grow and what to stop.

How we work

A standing engagement has a simple rhythm: learn, act, assess.

We repeat the cycle, updating priorities as the company’s needs change and as we learn its processes.

  1. 01

    We learn the company

    We talk to the board and to the people who own the processes. We collect running tests, needs and ideas, and agree a shared starting point.

    What you get: A priority list, named owners, metrics and agreed next steps.

  2. 02

    We run the chosen initiatives

    We refine the scope of the chosen work, bring in the specialists needed and coordinate the work on both sides.

    What you get: Visible progress, clearly named obstacles and the list of decisions that unblock further work.

  3. 03

    We assess results and set the next step

    We sum up effects, risks and lessons. The board gets a recommendation and the priority list is updated for the next period.

    What you get: A documented decision and a plan for further work with a clear split of responsibility.

This is not a one-off project. With every cycle we know the company’s processes better and can make sharper decisions. Larger integrations, ground-up software, formal audits and ongoing maintenance require a separate scope.

What we deliver

A decision, a working solution and a handover to your company.

Material for a decision

A list of initiatives with cost, risk, metric and recommendation. The board gets a basis for a decision, not a presentation about AI.

Concretely: a monthly summary, a decision and risk register, and the rationale for buying or building.

A working solution

An automation, integration, prototype, configuration or pilot, launched and tested in your company within the team’s monthly pool of hours.

Concretely: at a law firm this was daily automatic report retrieval without a standard integration, plus monitoring that chases missing entries by itself.

Handover to the company

A short manual, a named process owner and a way to measure. The solution has to work when we are not there.

Concretely: working documentation, a trained owner on your side, an agreed metric and the threshold that calls for action.

Case study

See what a real AI implementation looks like.

We describe the starting point, the change we made and the measurable effect, with no shortcuts and no invented examples.

  • Result of the engagement
  • Legal services

A law firm regained control of its team’s work and of 50 retainer clients

Monitoring detects scope overruns, extra work to be billed and clients who need contact or closer attention.

Key result

Regular reporting gaps cut by over 80% after a simple automation supervised by dyrektor.ai.

Go to case studies

Terms

A monthly subscription with an agreed pool of hours from the whole team.

You are not buying meetings or advisor hours. You are buying the team’s availability, the lead person and the specialists, which we spend where it creates the most value that month.

Billing
A monthly subscription paid in advance, covering an agreed pool of hours of the team’s work.
What the pool covers
The lead person’s and the team’s work: analysis, meetings, preparation, implementation, tests, documentation and coordination.
Extra hours
Only after your prior approval, at an agreed hourly rate. Never without consent.
Outside the subscription

Priced separately, before we start: dedicated systems and websites · large integrations, migrations and ongoing maintenance · penetration tests and formal audits · extended legal support and high-risk AI systems · guaranteed response times and 24/7 availability.

Tools, licences and infrastructure are paid by the company.

People and standard

Real people stand behind the service, working to one shared method.

Rafał and Michał created dyrektor.ai and develop the standard of the service. The engagement does not depend on one person being available: everyone works to the same method and draws on the right specialists.

Rafał Garbacz

Rafał Garbacz

He combines the perspective of an entrepreneur, an adviser and a practitioner. At dyrektor.ai he is responsible for contact with the board, priorities and running the engagement. He is the person a company comes to with a problem and holds to account for the result.

Michał Stanisławek

Michał Stanisławek

Michał has led technology teams and projects for over 20 years, from content management and television playout systems to Conversational AI and LLMs. At dyrektor.ai he is responsible for feasibility, architecture, measurement and the move from a decision to a working pilot.

Patryk Bajer

Patryk Bajer

Patryk has built software for over 20 years. Since 2019 he has been co-founder and CTO of utter.one, where he builds voice systems, chatbots and generative AI solutions, including the open-source Bonsai platform. At dyrektor.ai he is responsible for feasibility, architecture and measurement of proposed solutions.

Experience

Rafał Garbacz

He combines the perspective of an entrepreneur, an adviser and a practitioner. Since 2013 he has worked in technology, running a company and projects where technology had to turn into a business decision rather than a recommendation. At dyrektor.ai he is responsible for contact with the board, priorities and running the engagement. He is the person a company comes to with a problem and holds to account for the result.

  • in technology since 2013
  • TEDx speaker and co-creator of VoiceFest
  • president of the Instytut 42 Foundation and creator of Civil42

Experience

Michał Stanisławek

Michał has led technology teams and projects for over 20 years, from content management and television playout systems to Conversational AI and LLMs. He was responsible for the technical side of large transformations, built platforms running in environments where downtime immediately becomes a business problem, and led distributed teams in Poland, the United Kingdom and Australia. As founder of utter.one he has been developing and deploying AI solutions since 2018. W dyrektor.ai odpowiada za ocenę wykonalności, architekturę, pomiar oraz przejście od decyzji do działającego pilotażu.

  • over 20 years leading technology and teams
  • CTO of MediaAtmos and founder of utter.one
  • experience from mission-critical systems to production AI deployments

Experience

Patryk Bajer

Patryk has built software for over 20 years. He started with a student collaboration with Microsoft at Warsaw University of Technology, co-authoring Microsoft Poland’s App of the Year and winning the Polish edition of Imagine Cup. At Deluxe Entertainment Services Group he moved from senior engineer to software architect, responsible for systems running at international scale. Since 2019 he has been co-founder and CTO of utter.one, where he builds voice systems, chatbots and generative AI solutions, including the open-source Bonsai platform. At dyrektor.ai he is responsible for feasibility, architecture and measurement of proposed solutions.

  • over 20 years building production systems
  • Microsoft Poland App of the Year · Polish edition of Imagine Cup
  • 58 projects in 7 years, including the open-source Bonsai platform
The wider team
In the wider team we work with specialists from many fields, including lawyers, cybersecurity experts, designers, developers and many others.

Who it is for

This way of working fits a company ready to make decisions and act.

This engagement makes sense if

  • there are already AI ideas or first tests, but no shared direction and no ownership
  • at least one repeatable process has a result that can be measured
  • the board is ready to name a person accountable for the engagement and to make decisions
  • process owners can share knowledge and take part in the agreed work

Another form will suit you better if

  • you need round-the-clock technical support or full-time availability
  • you expect a large IT system to be built within the basic scope
  • the company cannot give access to people, data and decisions

Key questions

Before we talk.

We bill a monthly subscription covering an agreed pool of hours from the lead person and the team of specialists. The amount depends on the capacity we agree, which we set after the first call, once we know the scale of work your situation needs. Hours beyond the pool are billed only after your prior approval. Tools, licences and infrastructure are paid by the company.

First step

Let’s talk about one area that is stuck today.

We start with a 30-minute call with the owner, a board member or the person responsible for the area. No sales deck: we want to understand the situation and tell you plainly whether we can help.

Book a 30-minute call