# Brainhuggers Bureau — AI Venture Studio, Hamburg AI venture studio in Hamburg. Strategy, enablement and engineering for organizations putting AI to work: agentic systems, research agents, knowledge infrastructure, workflow automation. --- ## Navigation - METHOD - READING - FIELDWORK - START - [← THE OTHER SIDE] ## AI Venture Studio · Hamburg ### We build agentic systems and walk organizations through the transition. Strategy, enablement and engineering as one practice. Not three vendors. We run our own company on the systems we build for clients: autonomous research pipelines, agent-driven development, self-publishing infrastructure. What we recommend, we operate daily. - [ TAKE THE FIELD READING ↓ ] - [ BOOK 30 MINUTES ] ## We are our own proof no client logos · the systems we run ourselves - **96 PRs merged autonomously:** our own pipeline · ~10 weeks unsupervised - **13 DSP modules, in the browser:** Cortex · a synth you can play - **nightly gallery transmissions:** no human in the render loop - **20 yrs shipping inside agencies:** Fork · Jung von Matt · Valtech ## One system, three phases. *Awareness → Readiness → Steadiness* Most AI initiatives fail on operations, not models. The difference is whether there is a system underneath. ### Without a system - Shadow AI. Tools arrive through the back door, ungoverned. - Pilots that impress in the demo and die at go-live. - The capability walks out when the vendor does. ### With the Bureau - One shared target picture. Leadership decides from the same map. - Use cases in production, engineered where it counts. - The capability lives in your teams, not ours. ### 01 — Orientation & Target Picture *Awareness* Where AI actually stands, from generative to agentic, and what it means for your organization. From shared understanding to a first target picture. ### 02 — Competence & Personal Productivity *Readiness* Your people learn to work with AI in their daily routine. Faster, better, with judgment. Broad qualification plus hands-on depth where needed. ### 03 — Deep Implementation in Processes & Structures *Readiness → Steadiness* From the value chain to production systems: identify where AI pays off, model the target state with measurable value, build it in sprints. ### 04 — Sustainable Anchoring & Scale *Steadiness* The part most initiatives skip: keeping the capability alive after go-live, inside your own teams instead of your vendor. ### Module Index combinable · scoped to where you stand - **01 — AI Impulse & Target-Picture Workshop:** Executive session with live demos, then a working session with your leadership circle or AI board. - **02 — AI Foundations (e-learning):** Fundamentals, effective prompting, compliance. Self-paced for the whole workforce. - **02 — AI Assistant Training (on-site):** A full day in which every participant builds a working assistant for their own recurring tasks. - **03 — Value Chain & Use-Case Exploration:** Full-day workshop; prioritized use cases, preliminary roadmap and backlog. - **03 — Use Case Assessment:** Per use case: target process, architecture, success metrics, stack and data sourcing. - **03 — Use Case Development:** Sprint-based build to production readiness; low-code where sensible, engineered where it counts. - **03 — AI Infrastructure Concept:** Cloud vs. on-premise, platform and tool choice, privacy-compliant structures, pragmatic AI compliance guideline. - **04 — Guidance & Q&A Formats:** Moderated biweekly sessions: application experience, new tools, continuous impulses. - **04 — Advisory & Status Calls:** Sparring for the AI core team across the project lifetime. - **04 — Ambassador Program & Community:** Internal multipliers, jour fixes, guides and best practices. ## The Field Reading ### Where does your organization stand? Eight questions, four routes. It reads your shape, names the one route to start, and files a plate on the result. Nothing leaves your browser. The full diagnostic is available as a standalone tool: /field-reading ### Questions ### PEOPLE 1. **Do your people use the tools in daily work with judgment, or is it curiosity and copy-paste?** - Rare. A few individuals experiment alone. - Growing interest, no shared skill floor. - Most teams use it for real work. - A normal tool, used with judgment and shared standards. 2. **When the machine gets something wrong, do people catch it?** - We would not notice. - A few experts would. - Most teams review their own output. - Checking the machine is a normal habit. ### TECHNOLOGY 1. **Is there a sanctioned stack: models, an agent harness, compliant infrastructure?** - Shadow tools. Whatever people signed up for. - Some tooling, no standard. - A sanctioned stack is forming. - A governed stack, models and harness included. 2. **Can agents act inside your systems, or only chat?** - Chat only. Nothing is wired in. - A few manual, copy-paste integrations. - Agents act in some systems, with guardrails. - Agents act across systems, with permissions and review gates. ### PROCESS 1. **Have you re-drawn workflows for agents, or bolted AI onto processes built for humans?** - Bolted on. The process is unchanged. - Talking about it, nothing redrawn. - A few workflows redesigned around agents. - Processes are designed for human-agent teams. 2. **Is a use case actually in production, with owners and measured value?** - No. Pilots that never ship. - One, in a slow pilot. - A few in production. - In production, with owners and tracked value. ### THE SEAM 1. **Is the split between human and machine intentional and clear?** - Blurred. Nobody could tell you who did what. - Implicit. People sort of know. - Mostly clear for the work that matters. - Explicit. What the model does and what the human owns is named. 2. **Can you show your work? Is AI co-authorship visible and owned, not hidden?** - Hidden. We would not admit where AI was used. - Case by case, quietly. - Visible inside, not outside. - Visible and owned. The collaboration is a feature, not a secret. ### Readiness bands ### Awareness The picture is not shared yet. Start by getting leadership onto one map of what AI changes. Before tools, before pilots. ### Readiness The will is there; the system is not. The work now is competence, and putting real use cases into production with governance that holds. ### Steadiness You build and ship. The risk now is that the capability walks out with the vendor. Anchor it so it compounds inside your own teams. ### Recommended routes - **people:** Competence & Personal Productivity — AI Foundations + Assistant Training - **technology:** Deep Implementation in Processes & Structures — AI Infrastructure Concept - **process:** Deep Implementation in Processes & Structures — Value-Chain Exploration + Use Case Development - **seam:** Sustainable Anchoring & Scale — Operating-model design (the studio signature) ## Where to start, and what it takes *four shapes · combine or take one* Each shape maps to a phase. Pricing is scoped to the work, not sold off a shelf. **on request** ### 01 — Impulse Workshop *Awareness* A working session that turns scattered opinion into one target picture. ### 02 — Enablement Program *Readiness* Your workforce learns to work with AI, with judgment, in daily practice. ### 03 — Implementation Sprints *Readiness → Steadiness* Prioritized use cases taken to production in sprints, engineered where it counts. ### 04 — Anchoring Retainer *Steadiness* The capability kept alive after go-live, inside your teams instead of ours. ## Current Fieldwork ### What we're building right now ### File C-01 — Industrial Trade Strategic AI initiative. Competence building across the group. **Stack:** Workshops · enablement · roadmap **OUTCOME:** A shared target picture the group now plans against. ### File C-02 — Ad Sales Knowledge management with retrieval; back-office automation across JIRA, Outlook and ERP. **Stack:** RAGflow · LibreChat · Coolify **OUTCOME:** Institutional knowledge became findable. Status updates write themselves. ### File C-03 — Ad Sales Sales automation ending in rendered artifacts. Decks and motion assets, not dashboards. **Stack:** n8n · PowerPoint · nexrender **OUTCOME:** The team stopped rebuilding the same deck a hundred times. ### File C-04 — Finance Customised database fused with a workflow engine for diverse process handling. **Stack:** Database × workflow tooling **OUTCOME:** Every process got a home: states, owners, an audit trail. ### File C-05 — Sports Media AI-driven content platform for a league ecosystem, jointly with partners. **Stack:** Consortium build **OUTCOME:** A platform built with partners, not bought from a vendor. ### File C-06 — Institutions · AT Agentic ways of working for large institutions: target architectures, leadership simulations. **Stack:** With btveen, Vienna **OUTCOME:** Leadership felt delegation in simulation before buying it. ### File C-07 — E-commerce Product data made machine-readable so agents can find and act on the catalogue. **Stack:** Structured data · agent surfaces **OUTCOME:** The catalogue stopped being invisible to AI agents. *Client names withheld until cleared. References available in conversation.* ## Partners ### Who we build with ### PARTNER — Hamburg AI development and education. Joint delivery across strategy, enablement and engineering. One braid, no seams. ### PARTNER — Hamburg Business transformation advisory. Build, Grow, Transform. They shape the strategy, we build the agentic systems that carry it. Florian is the reason I joined Fork in 2006, and again in 2016. He brought Harway and Advanced Innovation in 2026. ### PARTNER — Vienna Agentic ways of working for institutions with real stakes, plus the btveen perspectives event series. ## Who operates this **Founder · operator** 20 years shipping code inside agencies. Fork, Jung von Matt, Valtech. Now a studio that operates its own systems in the open: a multi-agent development pipeline, a browser-native synth platform, a gallery that publishes itself nightly. All live, all ours. When we recommend a way of working, it is one we run ourselves. Daily. ## Common questions ### What does Brainhuggers Bureau do? Brainhuggers Bureau is an AI venture studio in Hamburg. We build agentic systems — autonomous research pipelines, agent-driven development, self-publishing infrastructure — and walk organizations through adopting them. Strategy, enablement and engineering as one practice, not three separate vendors. ### What is an AI venture studio? A venture studio builds companies and products in-house instead of only advising. As an AI venture studio, we run our own operations on the agentic systems we build, then help organizations put the same systems into production — so what we recommend, we already operate daily. ### How do you work with organizations? One system, three phases: Awareness (orientation and a shared target picture), Readiness (competence and hands-on productivity across your teams), and Steadiness (deep implementation and sustainable anchoring). The capability ends up living in your teams, not ours. ### What does an engagement cost, and how do we start? Engagements are scoped to where you stand — from a first orientation to deep implementation — with pricing on request. Start with a 30-minute call, or take the Field Reading diagnostic to find the thinnest route in. ### Where is Brainhuggers Bureau based? Hamburg, Germany. We work with organizations across the DACH region and beyond. ## Start with 30 minutes A short call to find out whether there is a fit. No obligations. **[ BOOK A 30-MIN SLOT ]** What we recommend, we operate daily. ## Footer - [Imprint](/imprint) - [Privacy](/privacy) - [The Bureau](/bureau) - [Patches](/patches) - [Signal Gallery](/gallery) - [[← BACK TO THE TRANSMISSION]](/)