Put AI to work.
Keep control.
yae sits between the AI your business uses and the systems it works in. You decide what each AI can see, what it can do, and when a person has to approve. yae checks every action before it runs, and keeps the record.
Start a pilotBring the AI you already use, through MCP or an API
Your AI can act. Who decides what it’s allowed to do?
Teams already use AI with customer and company data, and agents are starting to send, pay and change things. Access is usually all or nothing, and afterwards nobody can say who allowed what. yae gives every AI action one of three answers.
Clear rules
Give each AI exactly the access it needs, nothing more
Most AI gets far more access than it uses. With yae you set what each AI can do, what needs a person, and what it must never do. Anything you haven’t allowed is blocked, and you see exactly what changed before a new version goes live.
- Can, needs approval and never, set for each AI
- Choose which actions, on which systems, need a person
- Only admins can change rules, and every change is recorded
Bring your own AI
One set of rules for every AI you use
People switch between AI assistants like Claude and Gemini, and teams build their own agents. yae isn’t another AI. It sits beneath all of them, so the same rules apply whichever AI is asking. Before an AI acts, it asks yae. It can’t vouch for itself.
- No lock-in. Change AI providers without rewriting a single rule.
- Set once. A rule you write applies to every AI that can reach that system.
- Agents too. Your own agents and workflows check with yae the same way.
Systems and data
Connect the systems your AI works in
Connect the systems your AI touches and choose exactly which actions it may ask for. yae holds the keys, so an AI can’t reach a connected system without going through yae. Systems that support it also check yae’s one-time authorisation for each action.
- Any system with a web API: helpdesk, CRM, records or your own services
- yae holds the credentials, so your AI never sees them
- Each connection lists the exact actions AI may ask for
One place for all your AI
See every AI in your business, and stop any of them
Know which AIs are connected, which team they belong to and what they can reach. See exactly what a rule change adds or removes before it goes live, and stop any AI the moment something looks wrong.
- Every connected AI, its owner and its systems in one list
- See exactly what a rule change adds or removes before it goes live
- Stop an AI in one step; it stays stopped until an admin lets it back in
Approvals
A person approves the exact action
Not blanket access for the afternoon. The approver sees exactly what will happen, and only that happens.
Evidence
Know what every AI did, and why
Every request, decision, approval and outcome is recorded. Calm and readable day to day, with the full detail there when you need to prove what happened.
Connect your way
Works where your AI already runs
yae doesn’t replace your AI or rebuild your workflows. It sits at the step where an AI is about to do something, however that AI is built. Start with one action and grow from there.
Security
Built to hold when things go wrong
Most tools assume everything works. yae is designed for the day it doesn’t: an outage, a misbehaving AI, a missed payment.
Give AI access. Keep the authority.
Start with one team and one real workflow. Tell us which AI you use and what it should never do without asking, and we’ll set yae up with you.
Start a pilot