Why 90 Days is the Right Target
Ninety days is long enough to do something real and short enough to maintain organizational momentum. In our experience across 200+ enterprise Claude deployments, organizations that set a 90-day target reach measurable production results in that window more than 80% of the time. Organizations that plan for 6–12 months often find momentum stalls at Month 3 when early enthusiasm hasn't translated into visible results.
The 90-day roadmap is intentionally scoped to one department and 3–5 high-impact use cases — not to the entire organization. The goal at Day 90 is to have irrefutable proof that Claude delivers ROI in your specific context, building the business case for broader rollout. Think of it as a funded, accelerated pilot with production intent.
This roadmap assumes a mid-size enterprise (200–2,000 employees) adopting Claude.ai Teams or Enterprise. API integration projects follow a similar structure but with additional technical tracks. See our Implementation service for API-specific roadmaps.
Phase 1: Foundation (Weeks 1–3)
Phase 1 establishes the infrastructure, governance, and team structure before anyone touches Claude in a production context. Organizations that skip this phase spend Weeks 4–12 answering governance questions instead of deploying.
- Executive stakeholder alignment: confirm scope, department selection, success metrics, and executive sponsor
- Readiness assessment: IT infrastructure, data classification, security review
- Draft acceptable use policy with Legal review (use our Governance Framework template)
- License procurement: Claude.ai Teams or Enterprise — confirm seats and admin access
- Project team formation: Project Lead (0.5 FTE minimum), Department Champion, IT contact
- Admin workspace setup: Teams/Enterprise workspace configuration, SSO integration
- Data governance rules: what can and cannot be shared with Claude (PII, client data, etc.)
- Use case prioritization: select 3–5 highest-impact use cases for pilot department using time-savings analysis
- Prompt engineering: develop initial prompt templates for each selected use case
- Security & compliance sign-off: formal approval from CISO and Legal
- Select pilot cohort: 10–20 people from pilot department — mix of enthusiasts and skeptics
- Build role-specific training curriculum using our training curriculum framework
- Prepare hands-on exercises using real (anonymized) organizational documents
- Establish measurement baseline: time logs for target tasks, quality benchmarks
- Schedule all training sessions and communication cadence for the 90 days
Phase 2: Controlled Deployment (Weeks 4–8)
Phase 2 is where Claude goes live for the pilot cohort. The emphasis is on intensive support, daily feedback loops, and rapid iteration on the prompts and workflows developed in Phase 1.
- Foundation training: all pilot cohort members complete Foundation module (90 min)
- Role-specific training: department-specific module (2–3 hours) with real document exercises
- Day-1 access: every trained employee gets Claude access immediately after training
- Daily standup (15 min): what worked, what didn't, what questions do you have?
- Prompt refinement: update templates based on first-week user feedback
- Advanced module training: Claude Projects, system prompts, multi-step workflows
- Identify power users: who's using Claude most creatively? Document their workflows
- Integration review: can any Claude workflows integrate with existing tools (CRM, DMS, etc.)?
- First measurement checkpoint: 30-day active usage rate, self-reported time savings
- Address low-adoption users individually: what's blocking them?
- Mid-point metrics review: adoption rate, time savings, quality scores vs. targets
- Executive presentation: preliminary results, projected ROI, recommendation for Phase 3
- Workflow documentation: formalize the best Claude workflows discovered by power users
- Identify next cohort: if results are strong, begin planning Phase 3 expansion
- Course correction if needed: adjust curriculum, increase support, address systemic blockers
Phase 3: Production & Scale (Weeks 9–13)
Phase 3 takes the pilot from controlled experiment to production deployment for the initial department, while beginning planning for broader organizational rollout.
- Full department access: expand Claude access to all members of pilot department
- Advanced training for power users: MCP integrations (if applicable), Claude Code (for engineering), extended workflows
- Process formalization: Claude workflows incorporated into official team processes
- New hire onboarding integration: Claude training added to onboarding curriculum
- Peer advocacy: trained power users begin informal knowledge sharing
- Full 90-day ROI measurement: complete financial analysis, adoption rates, quality scores
- Executive presentation: business case for enterprise-wide rollout with projected ROI
- Phase 2 planning: identify next 2–3 departments, adjust timeline based on lessons learned
- Train-the-trainer programme: certify internal trainers for scaling (see our TTT guide)
- Documentation: implementation playbook for your organization — what worked, what to do differently
Implementation Team Structure
The minimum viable team for a 90-day Claude implementation:
- Executive Sponsor: C-suite or VP level. Provides air cover, removes blockers, signals organizational commitment. Time commitment: 2–3 hours per week for steering committee updates.
- Project Lead: The most critical role. Owns the implementation plan, coordinates across IT/Legal/HR/Department, tracks metrics, escalates blockers. Time commitment: 50%+ of working week for 90 days. This role cannot be done part-time and done well.
- Department Champion: Senior member of the pilot department who is an enthusiastic early adopter. Leads peer advocacy, surfaces workflow discoveries, provides ground-truth feedback on what's working. Time commitment: 25% of working week.
- IT Contact: Handles technical setup, security review, and integrations. Time commitment: intensive in Weeks 1–2, then as-needed support.
For organizations over 1,000 employees or API integration projects, add a dedicated Change Management Lead and Technical Architect. See our Implementation service for how we augment your internal team on larger deployments.