Ownership · Oct 2026 · 8 min read
The Handover Test
Ninety days after the vendor leaves, can your team run the system?
01
Thesis
A system is not finished when it launches. It is finished when the client can change it without asking the original builders to return.
We operated a multi-tenant intelligence platform for 3.5 years, then transitioned it to the client team that now runs and extends it.
EY IndiaThe client took the matching and scoring engine forward internally and productised the work as TalentMiner.
MOBOTIXThe platform is still operated by us at the client's request, which is a valid operating choice, not a technical lock-in.
02
The failure mode
Most delivery models optimise for acceptance, not ownership.
The cleanest question we know is simple: ninety days after the vendor leaves, can your team run the system? Not present it. Not open the dashboard. Run it. That means your team can deploy it, diagnose failures, change rules, trace data, answer why a number changed, and recover the system when a feed arrives malformed at 2 a.m.
Many enterprise systems are not built against that standard. They are built to pass acceptance, then quietly become permanent vendor dependence. The dependency is rarely announced. It hides in a proprietary runtime, an undocumented data model, a transformation only one engineer understands, or a workflow that was configured quickly and never explained.
This is why handover cannot be left until the final fortnight. By then the architecture has already made its decision. Either the system has been built around open, inspectable parts with the client team in mind, or it has not.
03
What to inspect
Ask for the model before you ask for the roadmap.
The model is where ownership starts. A vendor confident in the work should be able to show how the system represents customers, outlets, dealers, campaigns, orders, skills, documents, rules, events, or whatever objects your operation actually runs on. If the only artifact you can inspect is the interface, you are evaluating the surface while the machinery remains invisible.
The second artifact is the change path. Who changes a territory rule? Who adds a source? Who fixes an entity match? Who approves a new exception state? A system that needs the vendor for every operational change has not been transferred, even if the code repository has been handed over.
The third artifact is failure procedure. Production systems fail in ordinary ways: source schemas drift, credentials expire, queues fill, upstream teams send duplicates, users import yesterday's spreadsheet again. A clean handover includes the runbooks for these boring failures because those are the ones that decide whether the client can actually operate.
04
The better ending
There are two legitimate endings. Hidden dependence is not one of them.
One good ending is full client ownership. RYNO is the clearest example in our own history: we built and operated the platform long enough to stabilise it, then transitioned it to the client team. They now run and extend it. That is a handover.
Another good ending is a deliberate managed-operation model. MOBOTIX continues with us because that is the chosen operating arrangement. The difference is that this should be a commercial choice, not an architectural trap. The client should understand what is being operated, where the data lives, and what would be required to transition later.
The bad ending is ambiguity. Nobody knows what is proprietary, the internal team cannot tell whether a defect sits in data, logic, or interface, and every meaningful change becomes a new statement of work. That is not partnership. It is dependence discovered late.
05
Use this
Five questions before signing
Can we see the data model and the rule model?
Screens prove presentation. Models prove whether the system can be understood, audited, and changed.
Who can deploy, diagnose, and roll back the system?
If only the vendor can do it, the operational dependency is already present.
What is proprietary in the runtime?
Proprietary expertise is acceptable. Proprietary operational leverage is where lock-in usually lives.
What do we keep if the contract ends?
The answer should name code, data, documentation, infrastructure, credentials, and runbooks.
Can we talk to a client after transition?
Handover quality is easiest to judge when the vendor is no longer in the room.
Talk it through
Does your system look like this?
Fifteen minutes with an engineer, not sales. Bring the architecture, leave with a straight read on it.
Book a 15-min architecture call