MN Michael Ngangom

Implementation

Implementation is an operating system, not a project phase

By Michael NgangomJun 12, 20268 min read

Successful platforms connect solution design, delivery governance, adoption and a measurable path into operations.

I have seen technically sound banking programmes struggle after go-live because implementation was treated as a temporary delivery phase. The project plan finished; the operating model had barely started. My view, shaped by onboarding more than a hundred banking clients, is that implementation is the first version of how a platform will be run.

01Discover the real journey
02Prove the integration
03Rehearse production
04Operate and improve
The implementation loop I use: each stage must leave evidence that the next team can operate.

A go-live date is not a definition of done

A programme can be green on milestones and still be fragile. The API works, but nobody has agreed who owns a rejected payment at 02:00. The migration reconciles, but exception queues have no ageing threshold. The client has completed training, but the support team has never seen a realistic failure.

The discipline is to convert assumptions into named decisions, make readiness measurable, and keep implementation, support and product in the same conversation. I want the people who will carry the pager involved while the solution is still easy to change.

Design the handover while designing the solution

I prefer operational acceptance criteria alongside functional ones. Can we identify the customer impact? Is there an owner for every alert? Can a new analyst follow the runbook without calling the person who wrote it? Have we rehearsed rollback and reconciliation?

The best handover is not a final meeting or a large document. It is accumulated familiarity. Support joins design reviews, operations sees test evidence, and implementation stays through the first production cycles. That overlap looks expensive on a resource plan. It is usually cheaper than discovering a hidden dependency during the first serious incident.

Implementation should leave a system behind

A strong implementation produces more than a configured platform. It leaves decision logs, service measures, escalation paths, trained operators and a backlog informed by actual usage. Those assets become the operating system around the technology.

My test is simple: thirty days after go-live, can the service make routine decisions without the project team? If the answer is no, the project may have shipped, but implementation is not finished.