Tan Wei Ming · Head of Engineering · Logistics · SMB · 50 employees
Selasih's record shows no AI tooling inventory, no PR/review detail and no adoption numbers — the two biggest gaps are tooling (nothing indicates Cursor/Claude Code licences, shared IDE rules, MCP wiring or a secrets policy) and adoption (zero seat or usage signal across an engineering org that is simultaneously hiring aggressively on JobStreet). Day 2 lands flat without Day 1 groundwork, so the highest-leverage 90-day move is a sequenced rollout: a 30-minute tooling and velocity audit before the first cohort, then licences plus a shared rules repo and secret-scanning for one pilot squad, measured against a pre-workshop baseline of cycle time, PRs/dev/week and weekly active AI seats. Book one cohort of ~35 first and phase the rest once hiring settles.
CRM records no AI editor or agent stack for Selasih — no Cursor, Claude Code or Copilot licences, no MCP servers, no shared IDE rules and no secrets-handling note — and the products list is empty, so the stack is unknown at best. The score lands at 38 because Day 1's install-and-shared-rules modules would start from zero rather than from a partial setup.
No PR, CI or code-review data exists in the record and nothing suggests AI-assisted review, prompt hygiene or agent guardrails are in play. Day 2's PR-review and agentic-dev modules assume a baseline review workflow to attach to; the score of 32 reflects that this baseline is entirely unverified.
The only leadership signal is that the Head of Engineering is the named contact with a Mon 9:00am next step booked and account health at 58/100 — engagement without any demonstrated AI stance. The score of 58 reflects plausible but unproven sponsorship, with blast-radius tolerance for autonomous agents entirely unrecorded.
The single throughput signal is a JobStreet hiring spree four days ago, which points to a capacity shortfall rather than to flow data; no cycle time, PRs/dev/week or deploy frequency is recorded. The score of 42 reflects a real gap to close but no baseline against which to prove the workshop moved anything.
Zero licences, zero seats and zero usage figures appear anywhere in the record, and no sub-team distribution is known across the engineering org. The score of 22 is low because adoption is not merely weak here — it is unmeasured, the weakest possible starting point for a workshop priced per participant.
No defect, escaped-bug or time-to-feature data is recorded for Selasih; the only quality-adjacent signals are account health at 58 and a hiring spree that frequently follows bandwidth or retention strain. The score of 48 reflects outcome risk inferred from hiring pressure rather than from measured quality.
No proposals yet. The AI writes one from the readiness check.