How delivery runs

How a build runs, from scope to production

Every stage ends with something you can see: working software, a decision, or a check that has to pass.

Six stages from brief to production.

  1. 01Stage one

    Scope and acceptance criteria

    Agree the outcome, the priority workflow, the boundaries, and what has to be true for the work to be accepted.

  2. 02Week one

    First working walkthrough

    Working software in the first week, so the riskiest assumptions surface early.

  3. 03Core build

    The workflow and its integrations

    Build the agreed workflow, connect the named systems, demonstrate progress weekly.

  4. 04Beta

    Complete beta and instrumentation

    Finish the scoped beta, add the monitoring the system needs, prepare real acceptance scenarios.

  5. 05Acceptance

    Testing, security and performance checks

    Acceptance testing with you, plus the agreed critical-path, security, accessibility and performance checks.

  6. 06Launch

    Release and handover

    Ship to an agreed plan, confirm rollback, hand over code, credentials, documentation and training.

AI-assisted tools speed up parts of the work; they do not replace engineering accountability. A person owns architecture, review, security decisions and release approval on every project.

Timing is confirmed in writing once scope, dependencies and your inputs are understood. Larger or higher-risk programmes are phased rather than compressed into one release.

What working with us actually looks like.

How a PSOI Tech engagement runs, by decision point
Decision pointHow PSOI Tech works
Time to something you can useA working walkthrough in week one, not a status report
Visibility during the buildWeekly demonstrations of working software
What you get at the endDeployment, documentation, credentials and training — a handover another engineer could pick up
QualityTesting, security checks and human review before every release
After launchA written warranty period, then optional monthly support if you want it
If it is not a fitWe say so, and point you at a smaller first step or another provider

What happens after you send it

  1. 01

    A clear next step

    Whether this is a fit, which starting point suits it, and what to do first. If a smaller step is the more responsible route, we say so.

  2. 02

    Scope and a written quote

    The workflow in scope, the acceptance criteria and the exclusions — with price and payment schedule, all in writing before anything is built.

  3. 03

    Build, with weekly demonstrations

    Working software in the first week, and at every demonstration after it, so nothing is a surprise at handover.

Policies behind the process

Pricing, example scopes and the cost estimator

Delivery questions

01How quickly can a first release go live?

Most first releases are delivered in 2–6 weeks, depending on scope and integrations. That assumes a bounded first release and timely access to decision makers, content, data and any systems the build connects to. The first working walkthrough is targeted for week one, and the schedule is confirmed in writing after scoping. Larger or higher-risk products are phased rather than forced into an unsafe deadline.

02What types of projects are a good fit?

Internal tools, operations software, workflow automation, customer portals and bounded AI pilots — anything with a defined business problem, one primary workflow and an accountable decision maker.

03What projects are not a good fit?

Open-ended research, an unprioritised backlog, enterprise-wide transformation, safety-critical autonomy, work blocked on systems or data you cannot access, and any request for certification or guaranteed business results. Where that is the case we say so and point you at planning, a phased programme or another provider.

04How is AI used during development?

AI-assisted tools speed up research, implementation, testing and documentation. A person stays responsible for architecture, review, security decisions and release approval — consequential delivery decisions are not delegated to an autonomous system.

05Will client data be entered into AI tools?

Not by default. Data use, approved providers and access controls are agreed per engagement, and information is minimised to what the task needs. Regulated or confidential data is not sent to an AI service without an agreed basis, safeguards and your approval. Please keep secrets and sensitive personal data out of the initial brief.

06What happens after launch?

Handover covers deployment, repository access, credentials, documentation and training, and a written warranty covers qualifying defects for the period in the agreement. New features and ongoing operational support are a separate scope.

07Can you sign an NDA?

Yes, subject to reviewing the terms. An NDA sits alongside the project agreement rather than replacing it, or its privacy and security obligations.

Next step

Ready to put a real problem through this?

Describe the operating problem, the outcome you need and roughly when. You will receive a clear next step within 48 hours.