Data
Agents run on your data. We make what you have usable and safe, and we capture the judgement that was never written down.
Not analytics
This is not analytics, dashboards, a warehouse or BI. That is the old world, where data was for humans to look at in a report. In a team operating system, data is what the agents can see, reason over and remember.
Different job. Different bar. The agent has to act on it, not just chart it.
The two halves
01
Trapped and scattered across systems, email, docs, chat and people’s heads. Shaped for humans, not agents. Your data is not ready for agents just because it exists. We classify what can be used, what needs cleanup, what must stay out, and what needs human approval before it ever touches a workflow.
02
The most valuable input is usually what was never written down. Why the team does it this way, what good looks like, the judgement that lives in a few people’s heads. We capture it, so the knowledge that used to walk out the door becomes something the operating system keeps.
Built security first · the moat
What the agent sees. Where it goes. What leaks. What the regulator says. Whether your IP walks out the door. That fear stalls most enterprises, and it is exactly where we come from.
Most AI vendors come from technology. We come from cybersecurity. The Data pillar is built security first. We assume the worst case, we leave an evidence trail, and we make data safe to turn an agent loose on before we turn it loose.
Every agent action logged. Trust is an engineering problem, and we treat it like one.
For Australian regulated buyers, that means APRA CPS 234 and CPS 230 obligations and the Privacy Act sit inside the build, not bolted on after an incident.
The Map artifact · the data read
The Data read shows what exists, what is trapped, what is missing, what is sensitive, and what is safe to turn an agent loose on. It is the foundation every AI build stands on.