Platform
LiveVerifact
Verifiable legal knowledge for AI assistants. Every answer with an official citation and a signed proof — or an honest gap.
Project page verifact.network ↗
PioneerDesk GmbH · Deep Tech Integrator · Zangberg, Upper Bavaria
Deep tech integrator for sovereign
AI and knowledge systems.
We design, build and integrate systems that run on your infrastructure and can substantiate what they say — from the first assessment through to production. As proof, we run our own platforms, and they are live. Your knowledge stays in-house.
ISO 27001:2022 certified · Operated in Germany (STACKIT) or on your premises · Based in Upper Bavaria
Evidence, not claims
Four figures that describe our work.
6 systems
Our own platforms and tools in live operation — from legal verification to IT management.
As of September 2026 · /en/projects
ISO 27001:2022
0 non-conformities in the certification audit — externally certified.
DICIS · Reg. no. 250617 12 ISMS · valid until June 2028
USD 0
Paid to US hyperscalers. All our own systems run in Germany or locally on our own hardware.
STACKIT, IONOS, Infercom (LU/DE)
17 days
From whiteboard to production app — with two independent security reviews.
TMT MentalTech, 22 March – 7 April 2026
What we do
Assess. Build. Integrate.
A deep tech integrator sells neither a tool nor hours, but a system that runs in-house. We bring the building blocks together — open AI models, signed knowledge stores, line-of-business systems, the assistants your team already works with — and take responsibility for the result. EU requirements are built in from the start, not bolted on afterwards.
01 · Assess
Sovereignty Compass
One to two days with management and business units: which data must not leave the building, where does AI genuinely pay off, what is cheaper done with rules? The result is a prioritised map, not tool training. The fee is credited in full against a pilot.
02 · Build
Systems on your hardware
From a pilot on a single machine to a production system in your own data centre. The AI uses your company knowledge, substantiates its statements down to the source and logs every step traceably. What cannot be substantiated remains an honest gap.
03 · Integrate
Connected to where your team works
We plug the systems in where the work happens: in Claude, ChatGPT or Microsoft 365, in line-of-business systems and document stores — via open interfaces (MCP), with European inference partners where cloud is wanted. Your team is brought along, not overwhelmed.
What we have built
Our own systems as proof: built, operated, audited.
We sell no promises, only code that has run. Every one of these systems is live and operated by us — from signed legal verification to the AI workplace for companies. Plus tools for your own desk and openly documented research.
-
-
Platform
LiveSuveri
The sovereign AI workspace for companies in Europe. Open models in German data centres, every answer with model, cost and sources.
Project page suveri.de ↗ -
Platform
LiveKompetenzanker
AI literacy for the whole workforce — without jargon, with a credential that third parties can verify. EU-sovereign on STACKIT.
Project page kompetenzanker.de ↗ -
Platform
LiveOneLog
IT management with AI agents that resolve standard cases themselves — hosted in Germany, with a tamper-evident audit log for the evidence.
Project page onelog.pro ↗ -
Partner integration
LiveInfercom Router
EU-sovereign AI models directly in Claude, ChatGPT and the command line — one connector, one login, cost per call visible.
Project page suveri.de ↗ -
Platform
LiveTMT MentalTech
A coaching methodology as an app — to production in 17 days, without a single byte of client data being able to reach a US cloud.
Project page zalana-mentaltraining.de ↗
What sets us apart
Four pillars — and each one can be proven.
1 — Sovereign by Design
Your data stays in the EU or in your building and is not subject to US law (no CLOUD Act exposure). Operated in a German cloud or locally on your own hardware, European AI models, European inference partners — built that way from the start, not converted afterwards.
Evidence: Verifact, OneLog and Kompetenzanker on STACKIT · HUGIN and SmartScan fully local · Suveri with zero data retention.
2 — Evidence, not claims
Every one of our systems keeps a log that makes subsequent changes detectable: signed knowledge stores, chained audit trails, receipts that third parties can verify offline. And we are audited ourselves.
Evidence: Ed25519 receipts in Verifact · post-quantum-signed audit chains in OneLog, Kompetenzanker and SmartScan · our own ISO 27001:2022.
3 — An honest gap instead of a hallucination
Our AI systems say “there is no reliable basis for this” instead of inventing something plausible. That applies to answers for law firms just as it does to our own research projects, whose strongest result to date is a negative finding.
Evidence: verification threshold with abort in Verifact and HUGIN · refuted baseline figures in AeroForge · critic model in Suveri.
4 — Speed without sloppiness
17 days to a production app. A new platform with its own identity management set up in one day and hardened by four independent penetration reviews. For us, speed is the consequence of clean architecture, not its opposite.
Evidence: TMT MentalTech timeline 22 March – 7 April 2026 · Suveri security review 15 September 2026.
A closer look at our method: /en/craft
Going deeper
Sovereign IT management, NIS2 and the features behind them.
Specialist pages around OneLog and the security requirements for managed service providers — from the category and competitor comparisons to individual features. Product details and trials run via onelog.pro.
-
Category
Sovereign RMM
Remote management without US cloud — what “Made in Germany” means technically.
-
NIS2
NIS2 for managed service providers
Obligations, building blocks and evidence at a glance.
-
Comparison
OneLog vs. NinjaOne
The sovereign alternative to the US market leader — compared fairly.
-
Category
RMM for MSPs
Multi-tenancy, margin and an AI layer for managed service providers.
-
Product
Features
Apply updates automatically, detect unapproved software, resolve incidents autonomously, test changes on a pilot group first.
-
Insight
Data path check
Five questions to check whether your AI is giving your knowledge away.
Who we are
A deep tech integrator from Upper Bavaria. Not an agency.
Founded in 2024 by Marcus Lenczyk — over 25 years of enterprise IT in the DACH market, from server and network engineering through migrations on a scale of tens of thousands of clients to today's AI systems that take on tasks autonomously. Engagements at Siemens Energy, Deutsche Bank, Debeka, Sandoz and in KRITIS (German critical infrastructure) consulting under BSI guidelines.
We are small, focused and self-driven: we build what we recommend and run what we build. If we are not the right partner, we say so.
Key facts
- Registered office Zangberg, Upper Bavaria
- Place of jurisdiction AG Mühldorf / LG Traunstein
- Operations STACKIT (DE) · on-premise at the customer
- Audit standards ISO 27001:2022 · NIS2 · Art. 32 GDPR · BSI C5 (via STACKIT)
- Focus B2B · regulated industries in DACH and the EU
A project that has to stay in-house?
A 30-minute conversation. No sales phrases. We listen, assess and tell you honestly whether we are the right partner.