Skills application solution
Compare your project charter to this expert example for an ML-powered internal compliance chatbot. Use it to tighten stakeholder alignment, scope boundaries and measurable outcomes.

1. Project overview
An internal ML-powered chatbot supports employees with compliance questions 24/7 via the company portal. It improves policy access, reduces compliance team workload and scales support across departments.
2. Stakeholder summary
| Stakeholder group | Needs and expectations | Concerns or risks | Priority |
|---|---|---|---|
| Compliance | Accurate, current regulatory responses | Misinformation risk | High |
| HR | Onboarding support; easy to use | Overreliance; low adoption | Medium |
| IT | Integration, security, logging | Maintenance burden | High |
| End users | Simple answers; fast access | Confusion or low trust | High |
3. Project scope
ML objective: Develop a chatbot for common compliance queries, policy references and escalation of complex cases to humans.In scope
- FAQ-style answers across key compliance categories (data handling, ethics, reporting).
- Escalation of sensitive or complex queries to compliance officers.
- Integration with the internal compliance knowledge base and HR onboarding portal.
- Live policy updates through centralised knowledge management.
Out of scope
- External or customer-facing compliance queries.
- Legal interpretation beyond approved internal guidance.
- Third-party platforms outside the approved IT ecosystem.
Constraints and assumptions
- Content limited to pre-approved policy material from compliance.
- Deployment meets internal security standards and SSO.
- Launch before the next fiscal compliance audit cycle.
- Access to anonymised historical compliance queries for training.
4. Project timeline and milestones
| Phase | Description | Estimated duration | Key milestone |
|---|---|---|---|
| Data prep | Review, clean and categorise policies | Two weeks | Policy knowledge base reviewed |
| Build | MVP for high-frequency intents and FAQs | Three weeks | Chatbot MVP complete |
| Test | Evaluate on held-out compliance query set | Two weeks | Accuracy above 90% on sample queries |
| Launch | Pilot with three departments | Two weeks | Pilot begins |
| Iterate | Feedback, content and escalation updates | Two weeks | Feedback and update plan in place |
5. Roles and responsibilities
| Team member role | Area of responsibility | Name/function |
|---|---|---|
| Project manager | Timeline, stakeholder coordination | Name/role |
| Compliance SME | Policy curation; response validation | Name/role |
| Data scientist | Data pipeline, evaluation strategy | Name/role |
| ML engineer | Chatbot build and integration | Name/role |
| IT lead | Infrastructure, security, access | Name/role |
| HR liaison | Adoption and onboarding support | Name/role |
6. Success metrics and KPIs
| Metric | Target value | Notes |
|---|---|---|
| Chatbot usage rate | Greater than or equal to 70% in Q1 | Adoption across departments |
| Response accuracy | Greater than or equal to 90% on validation set | Measured during pilot |
| Escalation rate | Less than or equal to 15% | Share handled without human handoff |
| User satisfaction | Greater than or equal to 85% positive | Post-interaction survey |
| Ticket volume | At least 25% reduction within three months | Compliance team ticketing |
What this example does well
- Links stakeholder concerns to scope, milestones and metrics.
- Uses concrete, measurable outcomes tied to business impact.
- Keeps language clear for both technical and non-technical readers.
Tips for applying this skill in your role
- Surface stakeholder concerns early; they often hide constraints that should shape scope.
- Write the charter in plain language both sides can sign up to.
- Define metrics early so you collect the right telemetry from day one.
- Use milestones to pace communication under uncertainty.
- Treat the charter as a leadership artefact, not paperwork.
Action item: Reflection
What did you do well compared with this example?
Your reflection here...
Where could your scope or stakeholder mapping be more detailed or realistic?
Your reflection here...
What would you do differently in your own context?
Your reflection here...