Module 00 / Creative Human / Berlin
Ask why.
Creative Human builds AI systems that answer back. Every claim footnoted. Every run diagnosed. Every stakeholder askable. Every line of the engine public. Production AI from Berlin.
Ledger / System signature / 01
Agents in production
48
The full Brew Chai library, operating inside Creative Human's own engineering work today.
Levels of transparency
4
Claim, run, persona, engine. The trust ladder Glasshouse is built around.
Rehearsals on the record
5
Post-cutoff, real-world proof cases, each published on /glasshouse with sim output alongside documented ground truth.
01 / What we do
Creative Human builds AI-native systems that run in production. Glasshouse rehearses decisions against a simulated public — for announcement framings, crisis responses, national policy packages, and event-driven investment theses — and it answers the question "why?" at every level: claim, run, stakeholder, and engine. Brew Chai is the open-source orchestration engine underneath. Advisory and Executive Training are how we sit with the leadership teams adopting both. Berlin-rooted, European by posture, production-grade by commitment.
02 / The triad
Product
Glasshouse
Rehearses decisions against a simulated public before you make them. Every claim footnoted. Every run diagnosed. Every stakeholder askable. Every line of the engine public.
- — Rehearse announcements, crises, policy packages, repositionings, and event-driven decisions.
- — Ask any simulated stakeholder why they reacted the way they did.
- — Walk out with three formats: Markdown, PDF, and a board-ready slide deck.
Engine — open source
Brew Chai
The open-source orchestration engine we use to build Glasshouse — and everything else we ship. Apache 2.0 on GitHub.
- — 48 specialized agents, 35 workflows, one natural-language entrypoint.
- — Install from GitHub and read the code that built Glasshouse.
Service
Service
We sit with the leadership teams adopting AI as core infrastructure. Two formats.
- Strategic AI Advisory — recurring partnership for AI-mandated leadership.
- Executive AI Training — a structured cohort program for C-suite and boards.
03 / Proof
Selected rehearsals.
Five worked examples. We built each one to demonstrate what Glasshouse rehearses for a specific buyer archetype — F100 CEO, F500 CCO, special-situations fund MD, strategy-firm partner, ministry director. Every case is post-cutoff, rehearsed using only documents dated before the real event, and shown alongside what actually happened in public record.
Case Ref. 2025-01 / F100 CEO
Announcement Sim — Alphabet Q4 2024
Alphabet's $75B capex announcement framing erased ~$215B of market cap in the first 90 minutes. Meta announced a similar-sized package the same week with different framing and the stock rose. Same capex, different framing, opposite reaction.
Case Ref. 2025-02 / F500 CCO
Crisis Sim — UnitedHealth Q1 2025
The April 17, 2025 earnings release and guidance cut destroyed $115B of market cap in a single day. The compounding narrative arc over 26 days reached $280–300B before the CEO resigned.
Case Ref. 2025-03 / Special situations MD
Event-Driven Sim — Seven & i / Couche-Tard
Merger-arbitrage funds long Seven & i lost $90M in a single day when Couche-Tard withdrew its $47B bid. The stakeholder dynamics were public before the IC meeting. The desks that lost money had the information.
Case Ref. 2025-04 / Strategy-firm partner
B2B GTM Sim — Broadcom / VMware, March 2025
The 72-core minimum-purchase floor that leaked on 28 March 2025 was effectively retracted 14 days later. The rehearsal rebuilds the 14-day reversal sequence from the transformation history the commercial-software executive team demonstrably had in front of them on the morning of 27 March.
Case Ref. 2026-01 / Ministry director
Policy Sim — Kraftstoffmaßnahmenpaket / KPAnG
Germany's 12-Uhr-Regel fuel-price law, rehearsed before the law entered public debate, using only documents dated before 20 March 2026. What the simulation produced is shown alongside what ZDF, ADAC, and the Bundeskartellamt observed a week later.
System signature / Ledger note
We rehearsed our own positioning with Glasshouse before we shipped this site. The simulation told us to change a word. We did.S:dogfood/word-kill
04 / Location
From Berlin, for consequential decisions.
Creative Human is a Berlin-rooted engineering practice. We build AI-native systems for organizations across Europe and globally — and we built them from inside the European conversation about how AI should be governed, explained, and defended. Glasshouse, Brew Chai, and every decision rehearsal we run sits inside the tradition of Nachvollziehbarkeit — the right to understand what is being decided and how.
We're hiring — see open roles arrow_forward05 / Get in touch
Tell us what you're trying to decide.
Thank you — we have your message.
We read every inquiry and we respond in hours, not minutes. Someone from our team will be in touch at the address you provided.
If you'd like to explore what we do while you wait, see Glasshouse in action or read the engine on GitHub.
[email protected] · Berlin, Germany