Department Transformation Plans

One plan per department. Each plan translates the company-wide strategy (strategic transformation report) into that department’s operating reality at PRIME PRODUCTS — a deadline-driven ship supplier with bonded warehouses, tender business, and tens of thousands of SKUs.

How to read a department plan

Every plan uses the same section skeleton so plans aggregate cleanly into the program view:

1. Current state (assumptions to validate) — what the department does today, systems, pain points. Everything here is an input to discovery interviews (requirement analysis framework); validated findings become entries in the requirements register.
2. Target AI-first operating model — how the department works at M12.
3. AI use cases — the candidate list feeding the AI use-case pipeline.
4. Process transformation opportunities — redesign, not just assist.
5. Data sources / agents / automations / integrations — agents must match the AI agent registry; integrations reference M365 + SoftOne integrations and the architecture overview.
6. KPIs — aligned with the layers of the KPI framework. Baselines are captured in discovery; all show “TBD” until then.
7. Risks, training needs, deliverables, milestones — training aligns with the change management and training plan; milestones align with the 12-month roadmap.

The shared 12-month rhythm

All departments follow the same macro-cadence; the pilot month varies by wave.

• M1–M3 — Discovery: interviews, process mapping, data-source inventory, requirement capture. Every department goes through this.
• M4 — Platform ready: on-prem LLM assistant live for all ~50 daily users; baseline training.
• M5–M8 — Pilots: each department runs its lead use case as a pilot in its assigned wave (wave 1 = M5, wave 2 = M7, wave 3 = M9 start).
• M9–M12 — Rollout & agents: successful pilots become standard practice; approved agents go to production with human approval points; KPIs measured against M12 targets.

Wave assignment logic: wave 1 takes knowledge-heavy, low-operational-risk departments where a drafting assistant cannot break physical operations; wave 2 takes transaction-heavy departments once SoftOne read integration is stable; wave 3 takes departments whose use cases depend on wave-1/2 outputs or on sensitive data handling being proven.

Department summary

Department	Plan	Impact	Lead AI use case	Pilot wave
Executive Management	plan	High	KPI narrative briefing & board-pack drafting	Wave 2 (M7)
Sales	plan	High	RFQ-to-quote drafting from vessel requisitions	Wave 1 (M5)
Procurement	plan	High	Supplier quote comparison & PO drafting	Wave 2 (M7)
Logistics	plan	High	Delivery/berth coordination & customs-doc pre-fill	Wave 2 (M7)
Warehouse	plan	High	Pick-list optimization against sailing deadlines	Wave 3 (M9)
Finance	plan	High	Receivables-chasing drafts & cash-flow commentary	Wave 2 (M7)
Accounting	plan	Medium	Invoice/document matching assistant	Wave 3 (M9)
Human Resources	plan	Medium	Policy Q&A and onboarding assistant	Wave 3 (M9)
Customer Service	plan	High	Bilingual email triage & draft replies	Wave 1 (M5)
Marketing	plan	Medium	E-commerce product-content generation	Wave 3 (M9)
Technical Support	plan	Medium	Product/technical Q&A over manuals & datasheets	Wave 1 (M5)
IT	plan	High	Platform operations copilot & integration runbooks	Wave 2 (M7)
Compliance & Quality	plan	High	SOP drafting & audit-prep evidence collection	Wave 2 (M7)
Administration	plan	Medium	Correspondence drafting & document filing assistant	Wave 3 (M9)
Business Intelligence	plan	High	Natural-language querying over SoftOne	Wave 2 (M7)

Related

• 12-month roadmap — program-level view of the same waves.
• AI agent registry — canonical list of the agents named in these plans.
• Requirements register — where validated department needs are recorded as REQ-NNN.