Skip to content
Felix Schumann
Get in touch

Services / AI agents

An agent thatactually gets things done.

An agent needs more than a good prompt. I build specialist AI assistants with the right context, carefully scoped tools and an interface where people can review the results.

See it builtDiscuss your project

AI agents · Agent engineering · Tools and review points

Explore an example

Your task
Task01

A concrete task

  • 01Describe the goal
  • 02Provide context
  • 03Set boundaries

Your task shapes the implementation.

The everyday problem

An assistant that understands your working context.

An agent can research, bring information together and prepare a next step. It needs to know which sources to trust and which tools it may use. I make those boundaries part of the product so you can review its results and work with them.

Research, planning and preparation can be tailored to your working context. The steps an agent may perform on its own are explicitly defined. The model API and its handling of data are part of that decision.

What we will work on together.

The scope follows your project.
  1. 01

    A defined task

    A role, context sources and permissions for a specific use case

  2. 02

    The right tools

    Tool integrations, a conversation interface and reviewable results

  3. 03

    Tested against examples

    Tests with real examples, documented limits and operating costs

From my work

See what this looks like in practice.

All projects
In daily use2026

Content intelligence and production platform

A custom algorithm. Clear plans. Your content agent.

A custom-built algorithm, an editorial planner and an agent tailored to your content goals. From research to production.

  • A custom algorithm for your topics
  • A planning tool for the entire workflow
  • A content agent that fits your strategy
  • Next.js
  • TypeScript
  • MongoDB
  • +1
Explore the projectTwo editorial briefings per day
Live2026

AI workspace platform

Your AI tools. One workspace to build with them.

Claude Code, Codex and more AI tools in one workspace. Dedicated sessions and agent teams move projects forward in parallel.

  • Your AI tools, side by side
  • Build in parallel, stay in control
  • Collaboration with memory
  • Tauri 2
  • Rust
  • React 19
  • +3

Our first conversation

Which task would you give the agent?

You do not need a finished technical specification. Your idea, an example or a look at the current workflow is enough to start.

Let us talk it through
  1. 01

    What knowledge does the agent need, and which sources are authoritative?

  2. 02

    Should it suggest actions, read data or carry out approved steps?

  3. 03

    Which examples will we use to test quality, limits and failure modes?

You might also be wondering:

How do you handle data and external AI services?Do you work on confidential projects?What do you need for an initial enquiry?

Inside the implementation

MCP integration: Useful agents need understandable toolsTool names, inputs and outcomes: design an MCP integration that AI agents can use predictably and people can verify.AI content planning: Turn a pile of ideas into a defensible editorial choiceConnect audience needs, original expertise and production effort. Design transparent topic selection, editorial planning and approval states.Context engineering: The right information at the right momentWhat an agent needs to know now, what can be loaded later and why a longer context does not automatically produce better work.Human approvals: Where an AI agent should wait for youDesign approvals around concrete actions without overwhelming people with endless confirmation dialogs.RAG or fine-tuning: Which problem are you trying to solve?Retrieve current knowledge or improve repeated behavior: a practical decision guide and evaluation plan for internal AI assistants.Prompt injection: Why a good system prompt is not enoughHow external content can influence agents, and useful boundaries between reading, deciding and executing actions in an application.GPT-6 Astra: What the new model release means for real projectsA look back at the September 3 launch: a benchmark comparison, the limits of vendor results and a practical evaluation plan for adopting GPT-6 Astra.AI agent or AI platform? The distinction mattersWhy IXIOM is an AI working platform, what a specialist agent does and which questions should be answered before development starts.
Felix SchumannFelix SchumannAI automation engineer

Name it, I build it.

If something here looks like the problem you have, write to me. A short description of what is not working is enough to start.

Get in touch[email protected]