Service · Prague & remote

AI assistants that actually do the work

I build AI assistants and agents for businesses in Prague and across the EU: customer-support chatbots that answer from your own documents, internal assistants that search your knowledge base, and agentic workflows that research, write and process data. Recent work includes ANOTE, which turns doctors' voice recordings into structured medical reports, and multi-agent platforms for marketing content and market research.

Tools I use
01

What you get

Website chatbot

Answers customer questions from your documents 24/7 and hands over to a human when needed.

Knowledge-base assistant

Your team asks in plain language and gets answers with sources from internal documents (RAG).

AI agents and workflows

Multi-step agents that research, draft, check and file work — like the content pipeline behind pulkin.info.

Voice and document processing

Speech-to-text, OCR and data extraction that turn recordings and scans into structured data.

Connected to your systems

Assistants that read from and write to your CRM, e-mail, helpdesk or database.

Privacy by design

EU data storage, encryption and the right model for each task — as built for ANOTE's medical data.

02

Related projects

Real projects from my portfolio that this service builds on.

03

How we work together

  1. 01

    Free consultation

    A 30-minute call or meeting about your goals, customers and budget.

  2. 02

    Proposal & fixed quote

    A clear scope, timeline and price — before any work starts.

  3. 03

    Design & prototype

    You see and click through the solution early and shape it with feedback.

  4. 04

    Development & testing

    Built in short iterations with regular previews, tested on real devices.

  5. 05

    Launch & support

    Deployment, monitoring and help after launch — plus maintenance if you want it.

04

Frequently asked questions

Is our data safe with an AI assistant?

It can be. I design for privacy from the start: EU hosting, encryption, access rules and, where needed, models that run on your own infrastructure or on the device.

Which AI models do you use?

The one that fits the task — large models from providers such as OpenAI or Anthropic, or open-source models when data must stay in-house. The architecture lets you switch later.

How quickly can we try it?

A focused pilot on one use case, with your own data, can usually run within a few weeks. We measure it and only then decide whether to scale.

What does running an AI assistant cost?

Besides development there are usage costs for the model and hosting. I estimate them in the proposal and design the solution to keep them predictable.

06

Let's talk about your project.

Tell me what you need — a free 30-minute consultation, no obligation. I reply personally.