The handover: what happens when a process runs itself
The real outputs of automation beyond the software: documentation, dashboard, ownership, escalation, and maintenance, explained for operators tired of black-box promises.
You will never need to open the code.
That is not a promise about the quality of the engineering. It is how Nexibeo structures every automation build. The code runs. The process runs. Your job is to know what it does, how to check that it is still doing it, and who to call when something shifts. The handover is the part nobody talks about, because the chatbot demos look more exciting. But the handover is where the fixed fee stops being a project and starts being a quiet operational layer inside your company.
After a build, you receive four things. Each one answers a question you will have six months from now, on a Tuesday afternoon, when two people are out sick and the third just left for a competitor.
The four items you actually get
1. Process documentation built for a replacement hire
Most internal documentation is written by the person who built the thing. It skips the three steps they consider obvious. It uses abbreviations that made sense during the sprint. It assumes the reader has context.
You have seen this. You have handed a twenty-page PDF to a new operations lead and watched them spend the first two weeks decoding what page four actually means.
The documentation you receive is written for someone who joined the company last Monday. It maps the process from trigger to output, listing every tool touched, every decision branch, every fallback. It names the owner on your side, the owner on Nexibeo's side, and the exact URL where the process lives. It includes a one-page summary at the front, because the full document is for the week something breaks and the summary is for the board meeting where someone asks what the automation actually does.
No screenshots that will be outdated in three months. No internal jargon without a plain-language translation in parentheses. The format stays consistent across builds, so your team learns to read these documents once and then skims them in ten minutes for every future automation.
2. A status dashboard you check in under thirty seconds
You do not need a real-time command centre with pulsing red dots and eleven tabs. You need to know if the thing ran this morning and whether any outputs landed in the wrong place.
The dashboard you get shows volume, success rate, and the last completed run. If something fell into manual review, it shows the item and the reason. No alerts configured by default, because alerts without a response procedure are just noise. Nexibeo sets up the first alert only after you define who receives it, what they do upon seeing it, and what escalation looks like if they are unavailable.
Most dashboards inside companies are graveyards of good intentions. Someone built them during a data initiative. Nobody opens them. The URL expired when the BI tool subscription changed. The one you get at handover has a single link, a single page, and a single metric that tells you whether you need to act today. If you do not need to act, you close the tab.
3. A named owner on your side and a named contact on ours
Automation without ownership decays. Someone sets a new data source six months in. The field name changes slightly. The automation keeps running, but now it pulls a column that is always null. Nobody notices because the dashboard still shows green. Until a customer asks why the report they received contains blank sections.
You designate one person internally. That person does not maintain the code. They receive the documentation, the dashboard access, and a thirty-minute walkthrough during the handover call. Their job is to notice when the process's environment changes and to tell Nexibeo before the automation starts producing garbage.
On our side, you receive a direct contact. Not a ticket queue. Not a shared Slack channel monitored by five people who all assume someone else will reply. One person who already knows your process because they built or reviewed it. When you email them with a change, they acknowledge it within one business day and schedule the update.
This pair of owners is the difference between automation that lasts eighteen months and automation that silently fails in week six.
4. A written escalation path with time commitments
Severity levels are usually fiction. Every ticket is P1 when the sender is stressed.
Your handover includes three levels written in plain conditions. Level one: the automation missed a single run, outputs are recoverable. Response within one business day. Level two: the automation is producing incorrect outputs that are reaching customers or financial records. Response within four hours. Level three: the automation is down and a dependent process cannot run until it is restored. Response within one hour.
Each level names who on your team is authorised to declare it and who on our team responds. This is not a service-level agreement buried in a contract appendix. It is a single page kept in the same folder as the documentation. The operations lead can find it faster than they can find the old SLA from the CRM vendor.
What 'maintenance' covers for the next year
Maintenance in software usually means keeping the servers on and applying security patches. For an automation that touches your actual business logic, that is not enough.
Your processes shift. A supplier changes the format of their weekly inventory file. A compliance requirement adds a new approval step. Your team reorganises and the person who used to review the flagged outputs moves to a different department. These are not bugs. They are normal. They will happen three to seven times in the first year.
Maintenance covers them. When the supplier file format changes, Nexibeo updates the parsing logic. When the approval step moves, we reroute the output. When the reviewer changes, we update the notification recipient and walk the new person through what they are looking at.
Maintenance also covers monitoring. Every automated process sits inside an environment. APIs get deprecated. Authentication tokens expire. Rate limits tighten. Nexibeo watches for these changes during scheduled monthly checks and fixes them before your team notices a failure.
If a third-party tool you use raises its prices or discontinues a feature the automation depends on, Nexibeo proposes a migration path. You approve it. We handle the switch. The documentation and dashboard are updated to reflect the new setup.
The line between maintenance and a new build is clear. If you want the automation to do something it was never designed to do, that is a new scope. If you want it to continue doing what it already does despite a change in its surroundings, that is maintenance. Most changes are the second kind. Most automation contracts treat them as the first and invoice accordingly.
The handover call nobody rushes
The handover is not a forwarded email with four attachments. It is a scheduled call, recorded, thirty to forty minutes. Your operations lead attends. Your named owner attends. The Nexibeo builder attends.
The builder shares their screen. They show the process running live. They open the dashboard and point to the number that matters. They show where the documentation lives and scroll to the one-page summary so everyone sees how short it is. They name the escalation levels and confirm who on your side can declare each one.
Then they ask your team to perform two actions while they watch. Open the dashboard from your own machine. Find the escalation document without help. These are the only two things your team will need to do unsupervised in the first month. If they can do both, the call ends.
After the call, you receive the recording, a timestamped index of what was covered, and a confirmation that the maintenance period has started with an effective date. No further onboarding. No follow-up calls to explain what was already explained.
What you stop doing
This is the part the process documentation does not capture, because it describes what the automation does, not what it displaces.
Before the build, someone on your team spent part of every week moving data from one place to another. They exported a report, cleaned the columns, uploaded it to a portal, tagged the mismatches, emailed the summary to three people. None of this work appeared in their job description. It grew around a process that started small and became a fixture.
After the handover, that person stops doing those tasks. Their week gains back hours that are uneven at first. The first week, they check the dashboard four times because they do not trust it yet. By week four, they check it once and spend the reclaimed time on something that requires judgment.
The automation does not eliminate the role. It removes the part of the role that a script can do, so the person does the part that only a person can do. The handover makes that division explicit. The documentation says where the machine stops and the human starts. The dashboard shows whether the boundary is holding.
When the year ends
Twelve months after the handover, you decide whether to renew maintenance. By that point, you have lived through a supplier format change, a personnel switch, and at least one moment when the escalation path was tested. You know whether the process still matters to your operations. If it does, maintenance continues on the same terms. If the process has been retired or absorbed into another system, you give thirty days' notice and the automation is archived with a final documentation snapshot.
There is no renegotiation pressure. No automatic renewal buried in page seven of the contract. The decision is yours, informed by a year of the thing actually running.
Nexibeo builds and runs automations for companies that want the process handed over cleanly, not left as a prototype the original builder understood and nobody else can touch. The fixed fee covers the build, the handover, and the maintenance that keeps the process alive as your business shifts around it. If that sounds closer to what you actually need than another chatbot pilot, the door is open.
Have a process like the one above? Book a call.