Stop · Go · Build

To embrace this future, we have to stop in order to stabilise ourselves. Three urgent moves, starting now:

1. Agree a strategic halt to all legacy DTS feature development — a deliberate pause to remediate, stabilise and, where appropriate, take non-people cost out — savings we will need to fund the increased token usage that accelerated delivery demands.
2. Plan the UK + India future — design the accelerated delivery model across UK and India, spanning product, test and support, not just engineering.
3. Adopt & implement the RAPID thesis — our model for building at the new speed.

The Stop agenda

What we tackle when we stop:

• Client commitments — what have we promised? Forthcoming features and config work that don't pause just because we do.
• Tech debt — what must we remediate? Partial automated deployment, critical security gaps, deprecated components, thin test coverage.
• Infrastructure & cost — where can we optimise? Take cost out of the estate to help fund the AI tooling that faster delivery needs.
• Ways of working — where do we fall short? Processes, hand-offs and standards that can't keep up with the new pace.

The output: a prioritised remediation backlog and the savings to fund our AI tooling — the stable base everything next is built on.

The Go agenda

What it takes to implement RAPID:

• Tooling — which AI tools? The right tool for each step — the shared toolkit.
• Process — how do we work? One end-to-end flow, from definition to production.
• Gates — where are the checks? Quality and security gates built into the flow.
• Squads — who delivers it? The people and squad shapes across UK and India.

The shared toolkit

• Agents & platforms — IDEs, agents and platforms: chosen, configured and shared.
• Prompts — a curated, versioned library of proven prompts and patterns.
• Skills — repeatable capabilities we invoke, not reinvent.
• Playbooks — end-to-end ways of working across product, test and delivery.
• Standards & guardrails — quality, security and review gates that keep speed safe — the layer beneath the toolkit.

The next three months — three tracks in parallel

Track	Focus
STOP — pause & fix	Squads pause legacy DTS feature work to remediate tech debt, close security gaps and take cost out — building the stable base.
GO — build the engine	Squads stand up the RAPID engine — the shared toolkit, one end-to-end flow, quality gates and the UK + India squad shapes.
BUILD — prove it	A dedicated squad takes StudyBuddy from legacy to rebuild — the first product built at RAPID pace.

We build StudyBuddy to define the RAPID process — then apply that pattern to a bigger challenge: CXsmart.

The ask

Two decisions:

• Approve the stop — a strategic halt to legacy DTS feature development; a deliberate pause to remediate, stabilise and take cost out, with the savings funding our AI tooling. The green light to stand still, so we can move faster.
• Empowerment over ambiguity — freedom to make decisions at pace while the model evolves, without waiting for perfect certainty. Clarity will come from moving, not from planning.