
Launch is not the finish line. A business system lives or dies in its first twelve months in production, and the software maintenance contract you sign before go-live decides how those months go. Get the scope, the response rules and the exit terms right, and the platform keeps paying for itself. Get them wrong, and you inherit a system nobody owns.
This guide is for CIOs, heads of IT operations and procurement teams in the UAE who are about to sign one. It covers what the contract should include, how a software support SLA should work, the red flags to look for and the questions worth asking.
Testing covers the paths you planned. Production finds the ones you didn't. In the first year, four things happen to almost every system:
None of this is a failure. It is what running software looks like. The question is whether your contract has a home for each of these before they happen.
Building is a project: a fixed scope, a deadline and an acceptance test. Running is a service: ongoing, and measured by availability, response and how safely changes reach production.
Many vendors price the first carefully and treat the second as an afterthought. Ask for both in writing before the build starts. When the people who will support the system have a say in how it is built, you get the logging, monitoring, documentation and deployment pipeline that make post-launch support fast and cheap.
A good contract names each area of work, who does it and how you will know it was done. Use this table as a checklist against any proposal you receive.
| Area | What good looks like |
|---|---|
| Monitoring | Errors and slowdowns tracked with tools such as Sentry; alerts reach a named person before users notice |
| Incident response | Severity levels defined in writing, with agreed response and resolution targets for each |
| Change requests | Each request scoped and priced before work starts, then released on a regular schedule |
| Security patching | Framework, server and library patches applied on a stated cadence; critical fixes fast-tracked |
| Dependency updates | Libraries and runtimes kept on supported versions, not left until something breaks |
| Backups and restore tests | Backups automated, and restores tested on a schedule with the results reported |
| Access reviews | Access to production, code repositories (GitHub, Bitbucket) and cloud accounts (such as AWS) reviewed regularly; leavers removed the same day |
| Reporting | A monthly report: incidents, changes released, open risks and usage trends (for example from PostHog) |
| Handover and documentation | Architecture notes, runbooks and deployment steps kept current, so another team could take over |
| Exit terms | Code, data, credentials and documentation handed over on exit, with a defined transition period |
If a proposal is silent on any row, ask. Silence usually means the work is not included, or it will be billed later.
A software support SLA turns "we'll look after it" into commitments you can measure. Its core is a short list of severity levels, each with a plain definition:
For each level, agree a response target (when someone starts work) and a resolution target (when it is fixed or safely worked around). Response and resolution are different promises, so make sure the contract states both.
Agree the hours the targets apply. A staff portal used during the working week rarely needs round-the-clock cover; a payment flow that families or customers use at night does. Agree, too, how issues are reported (a portal, email or phone) and how performance against the targets is reported back to you each month.
Any one of these is worth a conversation before you sign. Three or more, and you are buying a build with no plan to run it.
Usually, yes. The builders know the code, the integrations and the reasons behind each decision, so fixes reach production faster and with less risk.
The trade-off is lock-in, and you manage it through the contract, not by splitting the work. Keep the code, cloud accounts and documentation in your name, and insist on exit terms. Then the team stays because the service is good, not because leaving is hard.
This is how we work with Emirates National Schools. We have worked alongside the ENS team since 2021 on admissions and fee payments, and have supported the platform since launch. Read more about our approach to support and maintenance.
What is the difference between application support and maintenance? Support responds to issues your users raise. Maintenance is the proactive work: monitoring, patches, updates and backups. A good contract covers both.
Who should own the code? You should. The repository, cloud accounts and domains belong in your organisation's name, with the vendor given access to do the work.
Can a new vendor take over a system they did not build? Yes, after a technical review of the code, hosting and documentation. Expect a short transition period while the new team learns the system.
How are changes priced under a maintenance contract? Agree in the contract which fixes are covered and how new work is handled. Each change request should be scoped and priced before work starts, never after.
The cheapest time to get post-launch support right is before the first line of code. If you are planning a launch, or already running a system without a clear support contract, book a discovery session. We will look at what you have and what a sensible maintenance scope looks like for it.

