Where the phone is the pipelineExplore more industriesWhat we owe you after the system goes live.
Every engagement includes an operate phase that starts the day your system goes live, and like every policy in our Trust Center, it is written down. This page sets out what it covers, how an incident is handled, what you get every month and how to leave. AI systems drift: models update, platforms change their APIs and your process moves. Operate is how a working system stays working.
Resolved · Northwind DentalCalendar access renewed. 2 held bookings written. Note added to the March report.
Alert · calendar sync failingRolled back to the last good version. Calls still answered, 2 bookings held for review.
Seven services. One table.
Everything the operate phase covers, how often it happens and what you see when it does. It is not a support contract. It is the part where the system keeps earning.
When something breaks at 3 AM.
Monitoring alerts a named owner on our side. Every automated action is logged, so we can see what happened rather than guess. Rollback is a documented procedure we test during phase three, not something we improvise during an incident.
- 01DetectA check fails or a number drifts outside its range.
- 02AlertThe named owner is told, with the log attached.
- 03Roll backThe last good version goes back in. Customers keep getting answered.
- 04FixThe cause is fixed and re-tested before it goes live again.
- 05ReportWhat happened and what changed, in writing and in the monthly report.
Response and resolution times depend on your hours, your channels and how critical the system is. Where your agreement sets them by incident type, they are written into it, and the monthly report measures every incident against them.
One report a month. Your number on page one.
The success metric is defined in week one, signed by a named person on your side, and printed on every monthly report. If a report cannot show it, something is wrong with the system or the measurement, and both are our problem.
- Your number against its targetThe one metric you signed, measured the same way every month
- Every alert and incidentWhat happened, how long it lasted and what changed
- What the sampling foundWrong answers, missed handovers and tone, with the fix for each
- The change logRecalibrations, integration fixes and rule changes, dated
- The next workflowThe queue, ordered by payback, for you to decide on
- Alerts
- 2, both resolved same night
- Conversations sampled
- 60, three tone fixes
- Changes
- Model update re-tested, calendar access renewed
- Next in the queue
- Quote follow-up, AmazePipeline
Month to month. No exit fee.
An operate contract you cannot leave is a confession that the work is not worth paying for voluntarily. So you can leave, and we make it easy.
The system goes live and operate starts. If it is not live by day 30 for a reason on our side, the build fee is refunded under our refund policy.
An initial term, because the first three months are where tuning actually happens on real traffic.
30 days notice, any month. No exit fee, no penalty, and a handover call with whoever comes next.
If you stop operate, you keep the system and it keeps running. You lose monitoring, drift correction and maintenance, so expect it to degrade over roughly six to twelve months as models and APIs change. Some clients accept that trade, particularly with a technical person in-house. We would rather you chose it knowingly.
Building the next system
That is a new engagement with its own scope, its own price and its own number. We will not quietly expand a retainer into a build. The queue tells you what is worth building next; you decide when.
How a new build runs- A named owner on your sideSomeone who can approve customer-facing changes and sign off the number.
- Access kept currentWhen a password, plan or platform changes, tell us before the system finds out.
- Approval on what customers hearEvery new customer-facing message goes past you before it goes live.
How our agents disclose, escalate and handle your data is in our AI use policy. Already a client with a question? Your named owner is your first contact, and every other way to reach us is on the contact page.
Before you sign. What owners ask.
The questions people ask about what happens after day 30.
Ask yours on a callMonitoring alerts a named owner on our side. Every automated action is logged, so we can see what happened rather than guess, and rollback is a documented procedure we tested during live traffic.
The first move is always to put the last good version back, so your customers keep getting answered while the cause is fixed.
No. A support contract waits for you to report a problem. Operate looks for drift before you notice it: monitoring, monthly sampling, recalibration when models update and maintenance when platforms change.
It is the part where the system keeps earning.
The build is a fixed-scope agreement. Operate is month to month after an initial 90 days, which exists because the first three months are where tuning actually happens.
After that, 30 days notice in any month, with no exit fee.
You keep the system and it keeps running. You lose monitoring, drift correction and maintenance, so it will degrade over roughly six to twelve months as models and APIs change.
We would rather you made that choice knowingly than found out later.
It is part of every engagement and priced with the build, after the Systems Diagnostic: $2,500, ten business days, credited in full against the build.
Telephony and model usage are passed through at cost, and you see the invoice.
Ask us what happens at 3 AM.
Bring the question to a 30 minute systems call. We will walk you through the alert, the rollback and the monthly report for your build, and tell you who is named on our side before you sign.
You are on the list.
We will email you times within one business day.
3 leaks found