A governed layer between
your systems and your decisions.
Every agent's authority is a setting you control. Every action is logged, and anything that moves money or stock waits for a human.
You choose how far
to trust each agent.
Every agent starts on Propose, and waits for your approval. Anything touching money or people is never automatic, at any level, for any customer. That is built into RetailAI, not a setting anyone can be talked out of. A master kill switch and a full audit trail sit behind every action.
The agent is switched off and does nothing.
Watches and raises a flag. Takes no action itself.
Drafts the action and waits for your approval. The default for every agent.
Acts automatically, up to a limit you set. Money and people never reach here.
Behind every decision
RetailAI makes.
Explainable
Every action carries the reason it was made.
Human oversight
A person is in the loop wherever the risk is real.
Reversible
Operational actions can be undone, not just logged.
Australian region.
Separated per client.
RetailAI runs on Google Cloud, in an Australian region, on infrastructure kept separate for each enterprise client. Controls are built and aligned to SOC 2, an independent security audit standard. Formal certification has not yet been completed, and we would rather say that plainly than let the words imply more than is true.
Everything you
need to know.
- Not yet. Controls are built and aligned to the SOC 2 standard, an independent security audit standard, and formal certification is in progress.