Background
Dominik Zachar has spent the last five years building systems where autonomous actors are trusted with real value: first in blockchain governance, now for AI agents.
As solution architect of the Summon Platform, the first DAO platform on Cardano, Dominik owned the architecture of its community, treasury and governance infrastructure: the first signature-based voting on Cardano, multisignature treasury wallets, a staking portal and vesting contracts. The platform grew from about 1,000 monthly users to about 100,000 concurrent participants during major votes, in an environment where every governance action moves real funds and nothing can be rolled back.
As CTO of Cypher Enterprises, the company that operated Summon, Dominik led the engineering team that delivered the governance portals Intersect used during Cardano’s transition to community governance: committee elections, chain-wide polls, ranked-choice voting and live tallies under load.
Alongside, Dominik led the development working group of ADAO, a community collective building open-source Cardano tooling, and helped lead early development of Round Table, the first multisignature wallet on Cardano. That work included advising NFT and DeFi teams and security audits of smart contracts and applications. Before blockchain: enterprise software engineering at CGI and Tieto, and Microsoft Dynamics CRM projects in the financial and automotive industries.
Why it carries over to AI agents
Governance is human oversight of autonomous systems. AI agents raise the questions DAOs and treasuries did: who may act, under which rules, with whose approval, within what limits, and how anyone can check afterwards what happened. The answer is the same too: give the actor capabilities, never the keys, and keep a record nobody can quietly edit.
At FeirAI
Dominik leads the design of FeirOS and of the rules each deployment runs under: what an agent may do, which actions need approval, how credentials stay out of the agent, and what the record must prove to an auditor. That includes Vultrino, the open-source FeirOS component that holds credentials so agents never see them.
Writing
Reference articles on governing AI agents in regulated operations:
- DORA regulation and AI agents: what the Head of Operations must be able to show
- EU AI Act in 2026: what changed for AI agent teams
- What is AI agent governance?
- AI agent audit trails: what proof actually requires
- Why AI agents should never hold credentials
- AI agent budgets and spending limits: why bounded overshoot is the honest model
- Guardrails vs. governance: what each one actually stops