Priya Devi Krishnan · Director of Engineering · AgriTech / B2B · SMB · 95 employees
Agromart Networks is a 95-person AgriTech/B2B company at Proposal stage with a Director of Engineering as the named buyer, two tagged champions, deal health 74 and a pricing call held today — so sponsorship is the strongest signal in the record. The two biggest gaps are the absence of any AI tooling or adoption baseline (no licence counts, MCP inventory, IDE rules or daily-usage data anywhere in the record) and no evidence of Day 2 preconditions, meaning PR-review AI integration, test gating and agent guardrails may not exist yet, so Day 2 could land above their level. The highest-leverage 90-day move is to attach a 30-minute pre-workshop tooling and usage audit to the RM 3,500/seat engagement, run one ~25-35 seat cohort, ship a shared rules file into their repos on Day 1, and re-measure adoption at day 30 and day 60.
The deal record carries no Cursor, Claude Code or Copilot seat count, no MCP server inventory and no IDE-rules or secrets-handling detail, so we cannot confirm the stack Day 1 depends on. The one concrete positive is authority: the contact is the Director of Engineering, so a single decision can standardise the editor stack across the org — which is why this sits at at_risk rather than critical.
Nothing in the supplied context shows AI in PR review, test generation, prompt hygiene or agent guardrails — the record only holds commercial fields (stage Proposal, 60% probability, pricing call today). With Day 2 preconditions unverified, the workflow-integration half of the workshop is the largest delivery risk and must be scoped against their actual repo before the cohort runs.
This is the strongest dimension: the record tags two champions, names the Director of Engineering personally on today's pricing call, and shows deal health 74 with a tomorrow-9am next step. That indicates leadership willing to sponsor and change workflows, but only one contact is named, so we cannot yet tell whether the second champion is a peer sponsor or just a logged tag.
No throughput baseline exists in the record — cycle time, PRs per developer per week, time-to-merge and deploy frequency are all absent, so the gap the workshop closes is currently undefined and unmeasurable at day 60. At 95 total employees the engineering group is small enough that one cohort likely covers it, making a pre-workshop baseline cheap to collect.
The record contains no licence utilisation figure, no daily/weekly AI usage percentage and no distribution across sub-teams, so we cannot tell whether this is a greenfield rollout or an underused-seat problem. The two-champion tag suggests grassroots interest exists, but interest is not the same as measured daily usage, and that gap is exactly what the pre-workshop survey must close.
No defect rate, escaped-bug count, time-to-feature or developer satisfaction data appears in the supplied context, so output quality is unscored rather than good or bad. The only quality-adjacent signal is deal health 74, a commercial indicator that tells us nothing about escaped bugs — hence a middling at_risk score that resets upward once a defect baseline is captured.
No proposals yet. The AI writes one from the readiness check.