Company

We build software that models how work is done.

Geist turns process evidence into maintained BPMN models, compares them with execution data, and supports reviewed changes.

Mission

One model of how the company really runs.

Companies depend on processes recorded across diagrams, systems, files, and staff knowledge. We are building one place to maintain the model, compare it with execution, and track improvements over time.

Vision

AI that understands the process, tracks it, and acts on it.

Our aim is AI that can read a process model, follow the digital parts of its execution, propose changes, and carry out approved actions with a clear audit trail.

The problem

Operational tools hold different parts of the process.

ERP, RPA, BPM, and AI agents each cover a different job. Their records and actions need to stay connected.

ERP

Connect records to the process.

ERP systems hold transactions and operational records. Understanding a handoff across systems also takes process context, ownership and the evidence around it.

RPA

Keep exceptions in view.

Automation handles defined steps. Teams still need to understand where exceptions go, who owns them and how a change affects the wider process.

BPM

Compare the model with the work.

Process models make responsibilities and decisions explicit. Keeping them useful means comparing the intended flow with execution data and updating them as work changes.

Agents

Give agents shared context.

An agent needs more than a task: the relevant process, evidence and permission to act. Those boundaries need to remain visible as work passes between people and agents.

Our thesis

Observe, model, act. Then do it again.

Operations teams often connect process diagrams, system data, recordings, ownership, and improvement work by hand. Geist connects those records to one maintained model and a controlled change process.

Michael Glück, founder

How we build

Principles, not slogans.

Build from observed work.

Use captured process evidence and execution data to create and update the model.

Useful defaults.

Start with clear defaults, then adjust them when the process requires more detail.

Built for daily work.

Dense, fast, and clear for people who manage operations every day.

Close the loop.

A finding that nobody acts on is noise. We measure ourselves by changes that ship and stick.

Team

Small, focused, building in Munich.

Geist is built by a small, focused team in Munich, led by founder M. Glück. We are deliberately lean: close to the product, close to the people who run real operations. If that sounds like your kind of place, we want to hear from you.

See open roles