Dynex.IT / Approach
Governed delivery, with evidence at every gate.
Large programmes rarely fail on technology. They fail on unclear scope, hidden risk and optimistic reporting. Our method is designed to surface all three early, while they are still cheap to fix.
Engagement models
Four ways to work with us.
We choose the commercial model with you, based on how well-defined the scope is and where the risk should sit.
| Model | Best when | How it works |
|---|---|---|
| Fixed-price delivery | Scope and outcomes can be defined up front. | An agreed price for an agreed outcome. We carry the delivery risk. Pay-for-success terms are available through Langford Hale. |
| Dedicated team | The roadmap will evolve and continuity matters. | A stable, cross-functional team working through your backlog, with capacity and cost reviewed monthly. |
| Systems-integrator partner | A larger programme needs a specialist workstream. | We deliver under the integrator's governance and standards, as a subcontractor or an embedded team. |
| Programme assurance | A programme is under way and needs an independent view. | A time-boxed health check, a written findings report and, if wanted, a recovery plan we help to deliver. |
Lifecycle
One lifecycle, four gates.
Each stage ends with a gate. A gate passes on evidence, such as a signed scope, a reviewed design, passing tests or a rehearsed cut-over, and never on a date alone.
Frame
Stakeholder mapping, business case, scope, risk register, integrated plan and commercial model.
Design
Requirements, architecture, security and data design, and a test strategy, all reviewed and recorded.
Build and assure
Short increments demonstrated to users, automated and independent testing, and user acceptance testing with the business.
Run and improve
Rehearsed release, hypercare, service-level-backed support and a prioritised improvement backlog.
Standards
Security and quality are built in, not bolted on.
Every engagement starts from the same baseline, and it rises to meet the client's own policies.
Least privilege
People, services and AI agents get only the access their role needs, and that access is reviewed.
Audit by default
Material actions are logged in append-only records, so the history can be shown rather than asserted.
Privacy by design
Data minimisation, residency and data-subject rights, aligned to POPIA and GDPR.
Secure pipeline
Code review, dependency and secret scanning, and environment promotion that cannot skip a gate.
Tested recovery
Backups are restored, not just taken. Disaster-recovery plans are rehearsed.
Honest reporting
Status is reported against evidence. Red is raised when it is first known, not when it becomes unavoidable.
AI in delivery
AI in our delivery, under the same governance.
We use governed AI agents across analysis, build, testing and operations. They work to the same gates as our people, within enforced permissions and budgets, and every action is recorded. Senior engineers direct the work and remain accountable for it.