Generative Analytics Use Cases with Joule, SAC and Datasphere
As of 2026-10-04
Generative analytics on SAP is no longer a slide: SAP Analytics Cloud ships Just Ask (natural-language questions on indexed models), Analytical Insights in Joule (the same JustAsk service behind Joule, with the user's identity propagated so SAC data access controls apply), AI-assisted commenting, chart summaries, calculations and data actions, and since QRC3 2026 an AI-assisted Prompt Insights widget; Joule in SAP Datasphere is generally available since Q1 2026; tabular foundation models (SAP-RPT-1.5, and since 15 September 2026 TabPFN-3.5 Plus) are callable from SAP AI Core. The senior skill is choosing two or three patterns where a real decision waits on a figure, proving the governed model underneath is ready, designing refusals, and measuring time-to-decision against a baseline captured before go-live — not promising "any question, any time".
What you will learn
- Map SAP's shipping generative analytics capabilities (Just Ask, Analytical Insights in Joule, AI-assisted SAC features, Prompt Insights widget, Joule in Datasphere, SAP-RPT-1.5 and TabPFN-3.5 Plus in SAP AI Core) to six business patterns
- Assess semantic-layer readiness (names, synonyms, types, hierarchies, access controls, owners) before committing to a pattern
- Configure and test one model for Just Ask and Joule with real user questions, synonyms, refusal cases and two access profiles
- Select two or three patterns in a workshop and define a time-to-decision baseline captured before go-live
Generative analytics means letting people ask, explain, narrate and predict on business data in natural language, with a language model or a foundation model doing part of the work and a governed semantic layer doing the arithmetic. On an SAP estate in September 2026 this is built from named, shipping capabilities — which is exactly why the consultant's job has moved from "can Joule do this?" to "which decision does this make faster, on which governed model, and how will we know?".
The SAP capability map, as of September 2026
Just Ask in SAP Analytics Cloud. Just Ask answers natural-language questions with charts and tables. Models are added and indexed through Manage Models; since the Q2 2024 update users, not only administrators, can index models. Administrators define synonyms for measures, dimensions and dimension values so that users can say "revenue" when the model says "net sales". Just Ask relies on correctly typed columns (numeric, text, date) and supports acquired models and SAP Datasphere models; the SAC help page Just Ask Restrictions lists what is not supported and must be read against your release before scoping.
Prerequisites
- Working knowledge of SAP Analytics Cloud stories and Datasphere analytic models
- Review core concepts first: C024, C017, C021
Outcomes
- Explain the Analytical Insights in Joule flow (JustAsk semantic search, query generation, HANA gateway, OAuth2 SAML bearer propagation) and why results respect SAC data access controls
- Choose between Smart Insights, Smart Predict, SAP-RPT-1.5 and TabPFN-3.5 Plus for an explanation or prediction need
- Write refusal requirements and test them alongside the happy path
- Produce a pattern-selection record with selected and rejected patterns, readiness gaps and metrics
Full module available to members. The full module adds: the decision framework · the end-to-end scenario walkthrough · the KPI scorecard · the anti-patterns · the code blocks · the knowledge check · the diagrams.