Named Role Assistants illustrate recurring enterprise work — not separate product SKUs. Each example maps a business role to governed tasks, outcomes, and evidence. Use these names when scoping pilots; see Best first use cases for industry defaults.
Example catalog
| Role Assistant | Typical role owner | First outcome focus | Related evaluation |
|---|---|---|---|
| Sales Assistant | Account executive, sales ops | CRM hygiene, pipeline review, meeting capture | CRM data quality |
| Customer Success Assistant | CSM, renewals | Health signals, renewal prep | Customer health review |
| Project Assistant | PMO, delivery lead | Status, risk, stakeholder updates | Project status review |
| Finance Assistant | AP/AR, controller | Invoice validation, exception triage | Invoice validation |
| Procurement Assistant | Procurement ops | Supplier alignment, PO cycle | Procurement cycle |
| Data Quality Assistant | Data steward | Master data alignment | Master data stewardship |
Sales Assistant
Supports pipeline and customer context work across CRM and documents — not open-ended research on ungoverned exports.
| Suitable when | Not suitable when |
|---|---|
| CRM is a certified datasource with role-scoped dimensions | CRM is disconnected and meetings never reach governed systems |
| Outcome is measurable (hygiene score, capture rate) | Success is defined only as "more chat usage" |
Links: Tasks and outcomes · Business outcome library
Customer Success Assistant
Focuses on retention and health using certified customer context — escalations stay with human authority.
| Suitable when | Not suitable when |
|---|---|
| Health metrics and contract context exist in connected systems | No governed view of customer entitlements or usage |
| Renewal playbooks require evidence of outreach | Assistant asked to bypass approval for contract changes |
Project Assistant
Coordinates initiative visibility across systems — status and risk surfacing, not autonomous project management.
| Suitable when | Not suitable when |
|---|---|
| Project and document systems share metadata relationships | No project entity or owner model in Enterprise Knowledge |
| Operators accept or correct task outcomes during pilot | Full autonomous resourcing is expected without gates |
Finance Assistant
Supports AP/AR exception review inside policy — not direct payment execution without approval.
| Suitable when | Not suitable when |
|---|---|
| Invoice and PO data are certified with dimensions | Finance data lives only in ad hoc spreadsheets |
| Audit expects evidence per reviewed exception | Regulatory scope requires human-only review with no AI assist |
Procurement Assistant
Aligns supplier and PO context for procurement ops — governed reads and suggested actions, not vendor portal credentials to models.
| Suitable when | Not suitable when |
|---|---|
| ERP or procurement system is certified | Supplier master data has no stewardship owner |
| First scope is validation, not autonomous PO creation | Policy forbids any AI touch on procurement records |
Data Quality Assistant
Helps stewards detect and resolve master data issues — reference alignment, duplicate signals, dimension gaps.
| Suitable when | Not suitable when |
|---|---|
| Master entities are modeled with dimensions | No authoritative system of record is connected |
| Stewardship roles are named before pilot | "Fix all data" is the only success criterion |
Limits
Examples describe planning patterns, not shipped assistant packages. Your first assistant depends on certified systems, dimensions, and a measurable outcome row in the Business outcome library.