From Smart Insights to Just Ask — Generative Analytics in SAC

What you will learn

• Classify any SAC feature as statistics/ML or generative AI in one sentence and state which one needs a data-protection review before enabling
• Apply the documented Smart Insights limits (top 5 dimensions, top 10 members, 200,000-node hierarchy fallback, 1,000,000-row ceiling) to decide whether a story question is feasible
• Scope a Smart Discovery analysis with Target, Entity and Measure and reject a use case that hits the multi-hierarchy or 1,000,000-cell restriction
• Prepare a Datasphere model for Just Ask (JustAskManage, Visibility, synonyms, query defaults, rules, per-user Datasphere access) and test it in Draft with a golden set
• Diagnose a failed Just Ask rollout from its symptoms (unsupported language, several models enabled, missing Datasphere access)
• Explain that Joule reaches into SAC through the Just Ask engine, and list four generative SAC features that may already be in the client licence

Module overview

Who this is for. You have delivered SAC stories with Smart Insights and Smart Discovery for years, and a client has just asked whether SAC "has generative AI like Joule now". The honest answer is "some of it, layered on top of features that were never generative". After this module (about 3 hours with a tenant open; SAC release QRC3 2026) you can sort every SAC "smart" feature into statistics or generative AI in one sentence, set up Just Ask for a Datasphere model and diagnose a failed rollout from its symptoms, and write a one-page scoping note that tells a controller which feature answers which question. Smart Predict is the statistical baseline you already know.

Prerequisites

• Working experience delivering SAC stories, including Smart Insights or Smart Discovery
• Module AI & LLM Fundamentals or equivalent working knowledge
• Modules “SAC Predictive Planning Hands-On” and “SAC Smart Predict Hands-On — Classification and Regression” recommended, so Smart Predict is understood as the statistical baseline this module builds away from

Outcomes

• Explain to a client, feature by feature, which parts of SAC's 'smart' surface are statistics/ML and which are genuinely generative AI.
• Configure Smart Insights exclusions and interpret its hierarchy and row-limit behaviour correctly.
• Scope a Smart Discovery use case against its Target/Entity/Measure model and documented restrictions.
• Diagnose a Just Ask rollout issue against its language support and Datasphere-access prerequisites, and explain the Joule/SAC integration direction correctly.