SaaSEye is a software company staffed by AI agents. Its departments, processes and engineering practices are written agreements that agents execute and named humans answer for.
What that means
Most companies keep their real operating rules in people's heads. Ours are documents — versioned, cross-referenced, and loaded directly by the agents that act on them. When a rule changes, the behaviour changes, because the document is the instruction.
That has an unusual consequence: the company can be audited the way software can. Every decision has a record, every agent action has a trail, and every agreement has a version history showing who changed it and why.
Departments, roles, engineering practice and commercial terms exist as agreements before any work depends on them.
Agents load the agreements that bind them and do the work — designing, building, testing, deploying, operating.
A named human is accountable for every department. Agents draft and build; people decide.
What we promise
Values are cheap to claim. Each of these is held up by a rule the company actually operates under, so you can check whether we mean it.
Work moves in small, shippable increments rather than long silent stretches. Because: small, reversible steps — and cheap to correct beats right the first time.
Testing is a separate function with its own authority, and agent-written work is verified by a model from a different vendor than produced it. Because: a model checking its own output shares the blind spot that produced the error.
Changing direction means changing a document, not renegotiating a shared understanding. Because: explicit over implicit.
We operate what we build, for the life of the engagement. Delivery is not the end of our involvement. Because: the human is accountable — and accountability that ends at handover is not accountability.
You are told what went wrong, including when it reflects badly on us. Because: truthful reporting over favourable reporting — our first operating rule, not our last.
How we are organised
Each department owns a set of decisions and has a named lead. The leads below are AI agents, each with a published role definition setting out what it may decide alone and what it must escalate.
Builds the solution. Owns how software is designed, written and reviewed.
Owns the standard a change must meet before it ships, and the gate that enforces it.
Owns how a change reaches a running system, and the isolation between customers.
Owns running systems: monitoring, incidents, and the standing operational duty.
Owns the agents themselves — how they are trusted, measured, and retired.
Accountable for the company. Ratifies every agreement the agents operate under.
On the names. The leads above are AI agents, not people. They carry human names because delegation needs an addressee and decisions need attribution — work is assigned to Mateo rather than to "the engineering agent". They are never presented as human beings, and any message from one identifies itself as an AI agent of SaaSEye. Further departments — design, security, data, support, sales, finance and legal — are defined and have leads appointed, but are not yet in service.
Plainly
SaaSEye was founded in August 2026 and is early. Rather than let a website imply otherwise, here is the state of it:
If that is the stage you want to engage a supplier at, we would like to hear from you. If it is not, we would rather you knew now.
Read by Jens Vedel Markussen, the accountable human — not by an agent, and not by a queue.