30 / 60 / 90 — how the engagement runs

You get a dollarized roadmap of where AI eliminates cost across finance, accounting and IR — even if you do nothing else with it.

• W1
  Immersion

  Embed with finance, accounting and IR. Shadow close, AP/AR, reporting cycles, board prep. Collect every template: board decks, investor reports, lender packages, deal memos, variance analyses. Confidential 30-min interviews — each team member, no managers present: “What takes the most time? What do you dread? What would you automate first?”
• W2
  Data flow mapping

  Trace lineage for each major process: source system → extraction → transformation → report. Flag manual re-keying, Excel-as-database, email workflows, version-control-by-filename. Map ERP, CRM, HRIS, banking, expense, BI. Document data quality issues.
• W3
  AI opportunity scoring & dollarization

  Score each process on the 6-dimension AI Readiness Scorecard (1–5 each). Prioritize by automation potential × business impact. For Wave 1 targets: current cost (hours × blended rate) vs target cost vs annualized savings. Tool choices per process (cloud vs local, model selection).
• W4
  Roadmap delivery

  Select 2–3 Wave 1 processes. Define scope, success criteria, build timeline, cost. Present findings + implementation plan to leadership. Sign SOW for Phase II if proceeding (fee credited if within 60 days).

I build the automation while running the finance function. Larger builds scale through experienced contract specialists I direct — you still manage one person.

• 1–2
  Pipeline build

  AI pipeline for first Wave 1 process: document ingestion → LLM extraction → structured output → validation gates → downstream system or report. Audit trail from day one. Validation layer cross-checks against system-of-record.
• 3–6
  Parallel run

  AI runs alongside the existing manual process for 2–4 cycles. Outputs compared. Accuracy measured, exceptions documented, prompts iterated. No cutover until accuracy is proven.
• 7–10
  Cutover & scale

  Cut over to AI-primary with human review. Manual process retired. Monitoring dashboards live (accuracy, exceptions, drift). Wave 2 discovery begins.
• 11–12
  Handoff & documentation

  Runbooks delivered. Finance, accounting and IR team trained on AI oversight: what to review, what to trust, what to escalate. Optional retainer for prompt/model updates.