The product was there and had been built with real pragmatism: a live platform, real users, AI features already in production. What did not hold was the environment underneath — designed to launch, not to absorb growing traffic and concurrent users.
The risk in these situations is not today's outage: it is reaching the point where every increase in usage forces a redesign, while nobody can say where the data goes or on what terms the service is provided.
Infrastructure
Migration onto something that scalesExporting the platform from its existing environment and configuring application, database, access, backups and deployment properly. The stated goal is absorbing growth without redesigning the system at every jump in usage.
Observability
Monitoring from day oneMonitoring set up straight away rather than after the first incident: knowing how the system is doing is part of the infrastructure, not an add-on.
Process
Faster development, better controlledOrganising the environment and the development process — repository, review, coding assistants — to move faster along a documented, traceable path. Speed and control are not in conflict when the process is set up properly.
Assessment
Data, AI and external vendorsA technical and organisational review of the AI features and data flows: what leaves the perimeter, towards which providers, and what that implies for GDPR, the AI Act and security.
Terms
Conditions of use and licensingTerms of use and licensing follow from the assessment, matching how the platform actually works. Ordinary matters are handled in-house; legal review is kept for the questions that deserve it.
This is the kind of work nobody sees from outside: no new features, no redesign. What changes is the trajectory — the platform can grow without being rebuilt, the team ships faster with more control, and the questions about data and terms have a written answer before anyone turns up to ask them.