The short version
- We tell you the truth on the call: if an agent is the wrong fit, you'll hear it before you spend anything.
- You decide how far it's ever trusted, on a five-step control ramp you set and can stop with one tap.
- Every tool is granted permission by permission, so a mistake stays small and reversible.
- If the agent we set up doesn't deliver, there's a guarantee behind it and no lock-in.
Behind every real decision about money sits one quiet question: what happens if I spend this and it doesn't deliver? Most people don't ask it on the call. This page answers it before you have to.
You are weighing up an AI agent for your business. That is a real commitment of money, time and trust, so it makes sense to want a clear view of the downside before you say yes. We keep that downside small in three ways: we tell you the truth about whether an agent fits, you decide how much it is ever allowed to do, and a guarantee stands behind the work.
We'll tell you the truth on the call
The first thing that protects you happens before you have paid a cent. On the scoping call, we look at the actual work you want handled and tell you straight whether an AI agent is the right tool for it.
Sometimes it isn't. Some problems are better solved with a simple piece of software, a process change or a person. If that is your situation, we will say so and point you at the better answer. We would rather pass on the work than sell you something that won't deliver: a customer we set up wrong costs us more than a customer we never had.
If an agent is the wrong fit for your problem, you will hear it from us first, not find out three weeks in.
You never hand over more than you trust
The bigger worry is usually not the money. It is the idea of an AI system loose in your business, touching your inbox, your customers and your data, doing something you didn't sanction. So we built the whole thing around one rule: the agent only ever has the trust you have given it, and you give that trust one step at a time.
We call it the control ramp. Every agent starts at the bottom and only climbs as you decide you are comfortable.
Locked down
It starts on low-stakes work only, with no power to change anything that matters. You watch how it behaves before it touches anything real.
Reads
Next it can look at your tools and your data to understand the work, but it still cannot act. Reading is safe; nothing it does here can be seen by a customer.
Drafts for your sign-off
Now it prepares the work, the reply, the update, the message and hands it to you to approve. Nothing goes out without your yes. This is where most owners spend the early weeks.
Sends
Once you trust the drafts, you let it send the routine ones on its own and keep approving anything that is higher stakes. You move the line wherever you want it.
Runs on its own
For the work it has earned your trust on, it just handles it. You set this pace. There is no point on the ramp you are pushed past before you are ready.
This is permission by permission, not all or nothing. Each tool the agent connects to, your email, your calendar, your messaging, is granted on its own. You can let it run your calendar freely while it still only drafts emails for you to check. Nothing is bundled.
Always available
One-click kill switch. If anything ever feels wrong, you stop the agent instantly, from anywhere, on any device. It pauses the moment you ask. There is no off switch to go hunting for in a crisis: it is always one tap away.
Put together, this means a mistake is cheap and reversible. The agent only operates where you have let it, you see the work before it goes out for as long as you want to and you can halt the whole thing in a second. The worst case is small and easy to undo, by design.
If it doesn't deliver, you're covered
The honesty on the call and the control ramp are how we stop things going wrong. The guarantee is what stands behind it if they do.
If the agent we set up does not deliver, you are covered. We are confident enough in the work to put that in writing, because the people we set up well are the ones who stay, refer and grow with us. A customer who feels trapped is no use to anyone. You should never feel like the only way out is to keep paying for something that isn't working.
The guarantee
We work hard so it does deliver. And if it doesn't, you are not left carrying it.
Worries, answered
What if it makes a mistake?
It starts on low-stakes work, every tool is granted permission by permission, and you see its work before it goes out. If something still slips, the one-click kill switch stops it instantly. A mistake stays small and reversible, never a runaway.
What if I want out?
You own your data and your accounts the whole way through. There is no lock-in and nothing held hostage. You can leave whenever you want and take everything with you.
What if it's just not for me?
Then we will most likely have told you on the call, before you spent anything. We would rather point you somewhere better than set you up for a result we can both see won't land.
Still want to talk it through?
Book a scoping call. We will look at your actual work and tell you straight whether an agent fits, and how we would keep the risk small while you find out.
Book a 15-minute call →