Module 04 / About Creative Human
We build AI-native systems that run in production.
Glasshouse rehearses decisions against a simulated public. Brew Chai is the open-source orchestration engine underneath — the engine we used to build Glasshouse itself. Advisory and Executive Training are how we partner with leadership teams that want to operate AI at production scale. We are a Berlin engineering practice, European by posture, production-grade by commitment.
01 / Cultural position
Why Berlin, why Germany, why Europe.
Creative Human is a Berlin company for the same reason it builds the product it builds. The two are not separable.
Europe has the cultural vocabulary for answerable AI already. GDPR's right to explanation. The EU AI Act's transparency requirements. The long German tradition of Nachvollziehbarkeit — comprehensibility, traceability, the right to understand what is being decided about you and how. The Bundestag's wariness of unexplainable automated decisions. German and European enterprises, and specifically the senior decision-makers accountable to boards, regulators, parliamentary committees, and their own conscience for consequential decisions, are culturally primed for a product whose promise is "you can always ask why, and we will always have an answer."
The dominant American AI narrative of 2024 to 2026 is trust the black box. Trained on opaque corpora, deployed as opaque weights, justified with opaque benchmarks, sold on opaque demos. Our position is the opposite. You do not have to trust our systems. You can interrogate them. Every claim is footnoted, every run is diagnosed, every simulated stakeholder is askable, and the engine that built all of it is open. Ask why.
02 / Brand pillars
Four working principles.
The four things Creative Human stands for. Non-negotiables, not aspirations.
Pillar / 01
Answerable by design
Every output our systems produce can be interrogated — claim, run, persona, engine. Every piece of infrastructure we ship is built so the question "why?" has a real answer.
Pillar / 02
Production-grade
Built to run, not to demo. That means monitoring, auditability, cost control, error handling, and the unglamorous operational surface that makes a system actually land inside a client organisation.
Pillar / 03
Orchestrated, not single-agent
Our systems have multiple specialised agents with defined roles, quality gates, and state. A single-model conversational interface is not a Creative Human product. It is the demonstration you give before you build the real thing.
Pillar / 04
European by anchor, global by reach
We are Berlin-rooted, Germany-rooted, EU-rooted. The cultural position on answerable AI is European, and we hold it directly. Clients are served globally; the work is done from here.
System signature / Dogfood call-out
We rehearsed our own positioning with Glasshouse before we shipped it. The simulation told us to kill a word. We did. 1 Glasshouse flagged one word as high-risk across 4 of 14 personas — read as “unproven, early-stage, not yet at production scale” by Fortune 500 buyer archetypes. The rewrite purged the word. Click to open the source rehearsal.
The footnote chip above links to the source rehearsal. Clicking opens the Glasshouse rehearsal that produced this finding — the personas that flagged it, the reasoning each persona gave, and the diagnostic panel for the run.
03 / Where we are
Berlin, Germany.
We are based in Berlin. We chose Berlin for three reasons: the European engineering talent pool is the deepest we know, the regulatory neighborhood already speaks the language we market in, and the city has a tradition of building things that are meant to last. Visits and in-person sessions happen here. Remote work across the EU is part of how we operate.
04 / We're hiring