You stay in control. Here's exactly how.
What an AI employee is allowed to do, what always waits for a person, what it can and can't reach, and how you turn it off in one step.
- 01You approve the big decisionsPricing, contracts, payments and sensitive messages wait for a person.
- 02It only reaches what the job needsIts own isolated workstation. No shared passwords. No other clients.
- 03You can see it and stop itEvery action logged. Access revoked in one step, any time.
Some decisions always wait for a person.Written down before launch. Not up to the AI.
Every AI employee works from a written job description and a list of limits you approve. When a request crosses one of those limits, it stops, tells the customer when to expect an answer, and flags your named approver.
One person on your side owns approvals and answers them within one business day. That's a condition of the 90-Day Results Guarantee, because we can only prove work we're allowed to do.
Watch one call hit the gate, step by step- Pricing and quotesIt never guesses a number.
- ContractsNothing gets signed or agreed.
- Payments and refundsNo money moves on its say-so.
- Sensitive messagesComplaints, disputes, anything personal.
- Public publishingPosts and pages go out under the approval rules you set.
- Anything it can't undoDeleting records, cancelling jobs, changing accounts.
You can add to this list. You can't make it shorter than this.
Every AI employee gets its own workstation.
An isolated cloud computer, set up for one role in one business. It holds only the apps, files and logins that job needs, and nothing else.
It doesn't get every identity by default. Each one is created only when the job requires it. Pick one to see the rule.
Work email
Created when the role sends or answers email, like follow-ups on quotes. It's the AI employee's own address, never yours.
It can- Send follow-ups you approved
- Answer from approved replies
- Log every message to your CRM
- Quotes
- Complaints
- Anything sensitive
Seven controls on every AI employee.The same on one team or a full workforce.
- 01
Limits written down
A job description, approved sources and a list of what it may not do. It works inside them and nothing else.
- 02
Identities by need
A work email, a phone number or a login is created only when the job requires it.
- 03
No shared passwords
Your own logins are never handed around. It uses its own, so its access is separate from everyone else's.
- 04
No cross-client context
What it learns about your business stays in your business. It never sees another client's work.
- 05
Every action logged
Calls, messages, bookings and record changes are logged. Your monthly report is counted from them, never estimated.
- 06
Failure tested on purpose
Before launch we cut a connection on purpose. When something's down it saves the details, alerts you and retries. Nothing gets lost.
- 07
Revoke in one step
Turn off access any time. The log of what it did stays.
Your data, handled plainly.
What we do with it.
- Your customers' information belongs to you. We run the system that handles it, under your policies.
- Your data stays in your systems and on that AI employee's workstation. It's used to do your work, nothing else.
- Service providers we need to run the work, like phone, hosting and AI models, get only what that task needs.
- Ask what's stored and where, and we'll show you.
What we never do.
- Sell your data or your customers' data. To anyone.
- Use one client's information in another client's work.
- Hand your logins to an AI employee.
- Let it act where no one can see. Every action is logged.
AI is always disclosed
We don't pass an AI employee off as a person. "An AI employee for this role" means an AI-powered software service doing defined tasks, never a staff member.
Recording consent
Calls answered by an AI employee may be recorded and transcribed. Voice flows carry a recording notice written for the states you work in.
Regulated work
Healthcare, legal and finance work starts with the paid assessment. We scope the data and controls first, and put appropriate privacy, security, compliance and human-review controls in place before anything goes live.
No certification badges on this page. If your work needs a specific standard, ask on the review call and we'll tell you straight what we can and can't meet.
Control questions, answered straight.
Can an AI employee spend my money or sign anything?
Who can see my data?
Do you need my passwords?
How do I turn it off?
What happens when something breaks?
Do you work with healthcare, legal or finance businesses?
See where an AI employee fits, and where it doesn't.
Seven questions. See what missed work is costing you and the team that fixes it. If it fits, a 30-minute review with Tony, where you set the limits before anything is built.