Can AI do the work, not just talk about it?
AI agents that research, decide and act. From focused automations to systems that run entire operations. Built for production.
Most of them run unwatched
We’ve been building them for years, from small automations that take work off someone’s desk to multi-agent systems that run entire processes.
The difference is autonomy. An agent doesn’t just answer questions: it decides, uses tools and moves work forward. So it also has to catch mistakes and know when a human steps in.
Our clients hire us to build what gives them an edge. Confidentiality is part of the job. Publishing the details could give that edge away. We’ve delivered in regulated environments and on projects worth hundreds of millions.
What we've built
A client product connected to multiple data sources, including legacy systems never built for AI. Built for hundreds of thousands of users in a regulated domain.
Behind one interface sit a knowledge graph, databases, documents, user history and memory. The client’s own experts can steer its knowledge and behavior without a developer.
A planning product for a large SaaS platform.
The client describes the project in their own words. The system asks questions, weighs hundreds of factors and produces a complete plan in hours. Experts stay in control and can inspect or override every assumption.
An agentic content system producing tens of thousands of pages in a specialist domain.
Separate agents write, review and translate. Weak evidence takes a more cautious route, and rejected content goes back for revision before anything goes live.
Also built by 2x10
PM Agents
plan the work, assign it and follow up on it.
E-Mail Agent
reads incoming mail, decides what matters and routes it.
OCR Agent
turns handwriting into structured data, without guessing.
Copy Agent
multiple workflows draft, compare and check content before it goes out.
What makes them hold up
Three questions matter once an agent reaches production. None is about the model.
- What does it cost at scale?A design that ignores cost per run gets expensive just when it starts becoming useful.
- What happens when it’s wrong?Failures will happen. The system needs to catch them before the user does and alert someone when it cannot.
- Can your team change it?If only the people who built the agent can maintain it, you have a liability. We build so your own teams can work with it.
What we could build for you
- Automate recurring work that currently takes up your team’s time.
- Turn research, documents and data into reliable output at scale.
- Add AI to systems you already run without starting with a migration.
- Connect processes across teams and tools into one working system.
- Build something new where AI creates a capability that wasn’t practical before.
Under the hood
Most engagements start small: talking it through, a few weeks of engineering, something real you can judge. If it works, we take it further.If it doesn’t, you find out early
Chatbots
More than a chat window
View themeDocument AI
PDF to product
View themeData Pipelines
Built to scale
View themeN° 06 Contact
Tell me what you’re planning
No form, no ticket system. Email me directly and you get me, not an inbox someone else manages.
Moritz Miedler · Founder, replies personally