Skip to content
R-Labs Inquire

R-Labs / How we work

Four disciplined phases. One artefact at each hand-off.

Every engagement moves through the same short loop. Each phase produces something the client can hold in their hand, act on the same week, and keep after the engagement ends.

Context

Fast, quiet absorption. The problem as it’s lived, not as it’s pitched.

The opening week is read-mostly. Access to the operating environment, the documents that actually shape the team’s thinking, the constraints that don’t make the first slide, and — critically — the decision that would change if this engagement were right.

Inputs
Structured interviews. Read access to the relevant systems. Existing strategy docs, board reads, or technical artefacts.
Outputs
A written context map. A named decision. A short, ranked list of unknowns.
Signal of readiness for Phase 02
The client and R-Labs share one sentence that describes the decision ahead.

Diagnosis

A small number of precise findings. What is actually going on, how we know, and which problems are worth solving versus surviving.

Diagnosis is not an exhaustive audit. The discipline is in what gets left out. A good diagnosis fits on one page, is separable into decisions the client can make, and survives someone disagreeing with it.

Inputs
The Phase 01 context map. Light instrumentation where it helps. Sampled traces of real work in flight.
Outputs
A memo — typically 2-4 pages. Ranked findings. Findings that are not worth chasing, named explicitly.
Signal of readiness for Phase 03
The client picks the one or two findings they want us to act on.

Design

The intervention: systems, agentic workflow, prompt scaffolds, evaluations, or a memo. Shaped to the team that will live with it, not an idealised org on a whiteboard.

Design is deliberately narrow. A working loop, a named tool, a revised cadence — something that is usable next week rather than architecturally perfect in six months. Where AI is in the loop, the eval pass is included in the design, not bolted on after.

Inputs
The Phase 02 findings the client chose. Real team constraints — headcount, review cadence, risk tolerance.
Outputs
A working artefact. A prompt library or scaffolded agentic loop. A decision cadence. A revised team contract. A memo the board can read.
Signal of readiness for Phase 04
The team runs the new pattern under real load for at least one full cycle.

Refinement

Quiet iteration under load. What broke, what held, what the next decision looks like.

Refinement is where the engagement ends well. The practice steps back as the team takes ownership. Hand-off is intentional: a brief written record of the pattern, the eval loop, and the known gaps, so nothing depends on the engagement continuing.

Inputs
Production traces of the new loop. Eval results. Any friction the team surfaced.
Outputs
A revision of the Phase 03 artefact. A short hand-off note. A named next decision — which may or may not involve further work with R-Labs.
Signal of completion
The team runs the practice without us in the loop for a full cycle.

Engagement shape.

Engagements are time-boxed and scope-explicit. A short strategy or audit lands in 1-3 weeks. A workflow redesign or prompt-system build lands in 3-6 weeks. Decision support is hours-to-days and runs on demand.

Pricing is flat-fee per engagement, known before the work starts. No retainers dressed up as relationships, no hourly billing, no change-order games.

Illustrative

What an engagement looks like in practice.

Representative engagement — no client identifying detail. Timeline and deliverables are typical, not guaranteed.

  1. Week 0

    Intake. 45-minute conversation. One paragraph describing the decision at stake. Fit confirmed; scope written down.

  2. Week 1

    Context. Access granted. Three structured interviews. Context map drafted and shared. One sentence on the decision, agreed.

  3. Week 2

    Diagnosis. Two-page memo delivered. Three findings, ranked. One decision surfaced for the client to make.

  4. Weeks 3–4

    Design. One intervention built and tested inside the team’s real workflow. Eval pass included by construction.

  5. Week 5

    Refinement. One revision based on live use. Hand-off note written. The team runs the pattern alone for the next cycle.

  6. Week 6+

    Independence. R-Labs is out. Follow-up by request only. A named next decision is on the table — separate scope, separate decision.

Start a conversation

The first conversation is a reading of the problem, not a sales pitch.