AI · automation · data foundations
Less, but better.
AI, automation and digital foundations, designed with clarity and built to run in production.
Empes helps companies turn ideas into working systems, from AI agents and digital characters to process automation, data foundations and practical transformation work.
What the terminal on this page shows, in turn:
- # what the case studies record. $ ingestion-framework --status. sources onboarded 300+. data managed 70+ TB. onboarding weeks → days. $ reporting-pipeline --status. manual reporting 4-6 h/week → removed. reports update themselves. $ character-system --status. batches consistent.
- $ empes --about. company Empes. practice AI, automation, data. based in Warsaw, Poland. model strategy + implementation. motto less, but better.
- $ empes services. AI agents and assistants. Synthesis audit, then automation. Infrastructure data and cloud. Studio websites and launches. # details: /services/.
- $ empes approach. 1 understand how the work runs today. 2 simplify remove steps first. 3 build documented, handed over. # the system before the tool.
- $ ls notes/. clean-up-after-ai-agents/ note. handing-tasks-to-ai-agents/ lesson. keeping-ai-agents-under-control/ guide. metadata-driven-ingestion/ guide. private-ai-chat-for-a-company/ guide. # 8 notes published.
- $ book --call. length 30 min. bring one process. you learn whether automation changes it. → empes.it/contact/.
What we do
Four services, one way of working
Empes AI
AI that does a real job
We help companies turn AI into something practical: a consistent digital character, an internal assistant, a workflow-specific agent, or a custom solution built around a real business need.
About Empes AIEmpes Synthesis
Simpler ways of working, then automation
We look at how your business works today, find where time and energy are being wasted, and design simpler ways of working using the right mix of no-code tools and custom automation.
About Empes SynthesisEmpes Infrastructure
Foundations your reporting can stand on
We build the data and technology foundation behind the scenes, so your reporting, automation and digital transformation efforts have something reliable to stand on.
About Empes InfrastructureEmpes Studio
Websites, brands and launches that keep working
We design and build websites, shape the brand they carry and plan the campaign that launches them, with the sign-ups, admin panels and newsletters behind them built properly too.
About Empes StudioWork
Systems in production
Most clients are described, not named; the sites we built are named with the clients' agreement. Every result is the one the case study records.
Global pharmaceutical · 2025
Scaling enterprise data infrastructure: a 300+ source ingestion framework
A global pharmaceutical company had more than 300 data sources and over 70 TB of data. We built one framework that brings every source in the same way, on AWS, so a new source takes days instead of weeks.
300+ data sources on one framework, 70+ TB
Read the case studyMid-sized company · 2026
A private AI workspace for a company's staff: one chat, chosen models, costs that stay in bounds
Staff at a mid-sized company were using public AI assistants on their own accounts. We built the company its own AI workspace on Open WebUI and LiteLLM, on its own server: work login, models the company chose, budgets that hold, and help built in.
Read the case study
Sports club, amateur league · 2026
Liga Astra: a tennis league that publishes its own results
Klub Astra's amateur tennis league moved from WordPress, Google Sheets and embedded dashboards to a fast static site with a small admin panel, where every saved score updates the public tables and the club ranking.
about 2 min from a saved score to the public tables
Read the case studyApproach
Understand, simplify, build
Understand
We start with the real business context, not with tools.
→ a map of how the work happens today
Simplify
We reduce complexity, identify what matters and propose the clearest path forward.
→ a redesigned process and a plan
Build
We implement with the right mix of AI, automation, data and infrastructure.
→ a working system, documented and handed over
How we build
The system before the tool.
Each project on this site taught the same lesson from a different side. Consistency in AI output is a systems problem, not a prompting one. Automating a broken process makes it break faster. Scale comes from better abstractions, not bigger servers.
- Inputsstructured, versioned, reviewed
- Processsimplified before it is automated
- Pipelinesone framework, configured per source
- Failurealerting, so you know first
- Handoffdocumented, run by your team
Notes
The method, written down
Guides, lessons learned and open topics drawn from the work, with every illustrative number labelled. All notes
- Clean up after your AI agentsAI agents leave a working copy of the code behind for every task, and nothing removes it. The three checks we use to remove one safely, and why cleanup is part of done.
- We built an AI dispatcher, then bypassed it ourselvesWhether a team uses an AI agent depends less on how good the agent is than on what it costs to hand it a task. What we saw on day one, and what we changed.
- Keeping AI agents under control: a method for predictable agent workHow to keep AI agents' work safe and predictable: people start the work, tasks a machine can check, a budget for each task, guards the agent cannot bypass, and a review.
Book a 30-min scoping call
Bring one process. We will tell you whether automation changes it.