Resources

How we think about agentic AI.

Research papers, essays, and the team behind Mantix. The deeper context for what we’re building.

Insights

Latest from the founders.

Explore our thoughts on the economic logic and architectural layer defining the next era of enterprise AI.

Substack

AI Agents Never Graduate from Onboarding

They learn the manual. They never get to day two.

Read on Substack →
Substack

The Mediocrity Tax: Why the Economic Logic Behind Enterprise Software Has Finally Collapsed

The replacement is not a better suite. It is a different logic entirely.

Read on Substack →
Substack

The Layer Nobody Is Building

Every platform era has a piece of infrastructure that becomes obvious only after someone builds it. For agentic AI, that moment is now.

Read on Substack →

Research

The agentic blueprint for the enterprise.

Our founders have published two whitepapers on the agentic enterprise — one on the economic and architectural impact of autonomous AI agents, and one on the fundamental challenge of context alignment in agentic UI systems.

Whitepaper · arXiv:2602.21401 · February 2026

The Headless Firm: How AI Reshapes Enterprise Boundaries

01 — Coordination Collapse: AI drastically reduces integration overhead from exponential to linear, making modular, agentic delegation economically sustainable.
02 — Headless Architecture: A new enterprise equilibrium defined by a personalized generative UI, a standardized protocol waist, and a competitive market of execution agents.
03 — Outcome-Based Verification: Integration moves away from manual bilateral governance toward protocol-mediated tool access and outcome-based verification.
arXiv Abstract → Read PDF →
As heard on AI For The C Suite · “The Disappearing Middle” →

Vision Paper · cs.AI · February 2026

Position: Context Alignment Is a First-Class Learning Problem in Agentic UI

01 — The Context Gap: Larger LLM context windows alone do not solve agentic UI — agents must learn to synchronize their internal state with dynamic, partially observable interfaces.
02 — Temporal Drift: The gap between an agent’s world model and the actual UI state at execution time is a critical and underexplored failure mode.
03 — Context Alignment Framework: A four-layer taxonomy for how agents should learn to navigate and manipulate complex enterprise software.
Read Whitepaper →

Founders

The team behind Mantix.

Co-Founder & CEO

Dr. Sebastian Wieczorek

12+ yrs enterprise AI @ SAP VP AI Technology & ML Foundation PhD TU Berlin EC & German Govt AI advisor

Co-founded SAP’s first enterprise AI unit in 2014 — years before the market understood what that meant — and built the Leonardo ML Foundation, SAP’s AI platform serving thousands of global organizations. As VP AI Technology, he shipped production-grade AI at a scale few startups ever reach before founding their own company. A recognized authority on responsible AI governance: EU Commission Expert, German Enquete Commission on AI, and Bitkom AI Council member.

Ex-SAP VP AI Technology PhD TU Berlin EU Commission AI Expert German Enquete Commission AI Bitkom AI Council

Co-Founder & CTO

Dr. Tassilo Klein

10+ yrs AI Research @ SAP Director SAP ML Research PhD TU Munich · Postdoc MIT & Harvard ELLIS member · Enterprise AI Research

Director of Enterprise AI Research and former SAP Distinguished Scientist. Holds a PhD from TU Munich and conducted advanced AI research at Harvard and MIT. Expert in NLP, tabular data, and agentic AI architecture. Member of the European Laboratory for Learning and Intelligent Systems (ELLIS).

Ex-SAP Director of AI Research ELLIS Member PhD TU Munich Harvard & MIT Researcher NLP & Tabular Data Agentic AI