/14Department Playbooks
Ten department playbooks showing what to assist now, what to buy, what may be worth building, who should own it, and what to measure.
Leader
16 min
Open +
Department Playbooks
Ten department playbooks showing what to assist now, what to buy, what may be worth building, who should own it, and what to measure.
Leader
16 min
The company needs one AI strategy, but each department needs a translation into its own work. Start with workflows, not a company-wide list of approved features.
Leave knowing / Give every department a useful first move
Department playbook / 1
Customer and growth teams.
| Department | Assist now | Buy when | Build when | Owner / proof |
|---|---|---|---|---|
| Sales & Revenue | Research, call prep, follow-up, CRM hygiene | The workflow is standard and already lives in your CRM | Your sales motion, account data, routing, or offer logic is genuinely unique | Revenue leader / seller time, response speed, conversion, data completeness |
| Marketing & Growth | Research synthesis, briefs, variants, reporting narratives | You need commodity creation or optimization inside an existing channel | Your brand system, approval flow, data, or content engine is a differentiator | Marketing leader / cycle time, approved output, performance, rework |
| Customer Success | Case summaries, answer drafts, account-risk synthesis | Your support stack already has mature AI using the context you need | Resolution needs proprietary product data, unusual workflows, or cross-system actions | CS leader / resolution time, escalation accuracy, satisfaction |
| Product Management | Research synthesis, feedback clustering, spec drafts | You mainly need meeting, research, or backlog assistance | Your product discovery evidence or decision workflow is a strategic capability | Product leader / learning cycle time, evidence coverage, decision speed |
| Design & UX | Research synthesis, content exploration, prototype assistance | You need commodity generation or prototyping | Your design system or production workflow can be encoded into a useful internal tool | Design leader / exploration time, rework, accessibility and quality review |
Department playbook / 2
Operating and technical teams.
| Department | Assist now | Buy when | Build when | Owner / proof |
|---|---|---|---|---|
| Business Operations | Intake, routing, document handling, exception summaries | The process is a standard back-office workflow | Your workflow crosses unusual systems, rules, approvals, or business logic | Ops leader / touches per case, cycle time, exception rate |
| People & Culture | Onboarding, policy explanations, role-material drafts | You need common HR assistance inside your existing HRIS | You have a distinctive internal knowledge or learning workflow with safe boundaries | People leader / admin time, completion, escalation and error rate |
| Finance | Extraction, variance narratives, scenario preparation | The work is standard close, AP, expense, or planning functionality | Your controls or planning logic require a narrow internal system with strong verification | Finance leader / review time, close cycle, exception accuracy |
| Data & Analytics | Query drafts, documentation, recurring analysis summaries | Your warehouse or BI product already provides governed assistance | The business needs a unique semantic, evidence, or decision layer across its data | Data leader / analyst cycle time, query quality, adoption, incorrect claims |
| Engineering | Tests, documentation, review assistance, migration drafts | You need strong coding assistance in normal repositories | You have proprietary developer workflows, internal platforms, or verified multi-step operations worth encoding | Engineering leader / cycle time, review burden, defects, incidents |
What every department owes the company
Local freedom needs a shared contract.
Use the same definition of approved data and prohibited data.
Name one accountable business owner for every live AI workflow.
Keep a simple inventory of what is being used and why.
Measure the before state before claiming the after state is better.
Share failed pilots so another department does not pay to relearn the same lesson.
Escalate common infrastructure needs instead of buying the same capability ten times.
Department workshop
Turn this into a 45-minute working session.
Try the template
We are mapping AI opportunities for this department: [DEPARTMENT] Its goals this quarter are: [GOALS] Its recurring workflows are: [WORKFLOWS] For each workflow, create a table with: 1. Current pain or delay 2. Best AI assistance opportunity 3. What must stay human 4. Buy / build / hybrid direction and why 5. Data and system access required 6. Business owner 7. One metric to prove value 8. One failure condition that would make us stop End with the two smallest pilots to run in the next 30 days. Do not recommend AI where normal software or a process fix is better.
Notes / Subscribe
Get the next note.
No schedule. No content calendar. A note goes out only when there is something worth understanding.