Joule in SAP Analytics Cloud: what it actually does today (and what it needs)

SAP's messaging in 2026 is almost entirely about Joule, Business AI and agents. If you work in SAP Analytics Cloud, the practical question is simpler: what does Joule actually do in SAC today, what do you need to switch it on, and what does it expect from your models? This guide answers that from SAP's own published material. We have not tested Joule hands-on; where sources differ, we say so.

The short answer: four capability types

SAP's reference architecture for Joule with SAP Analytics Cloud describes four kinds of capability:

• Analytical — ask questions about your data in natural language (for example "top 3 sales managers last month") and get a chart or table back.
• Informational — knowledge-based answers, such as how to do something in SAC.
• Functional — tasks carried out inside SAC stories.
• Scripting — generating scripts that add interactivity and automate behaviour in stories.

The analytical part is the one most people mean by "Joule in SAC", and it has a specific name in SAP's catalogue: Joule, Analytical Insights.

Analytical insights run on Just Ask

Joule doesn't query your data on its own. SAP describes analytical insights as integrating with Just Ask, SAC's natural-language query feature, and drawing on the models made available to it. That has a direct consequence: Joule can only answer from models that have been set up for Just Ask.

Independent write-ups describe the Just Ask setup in practice: a user with the Just Ask management permission adds a model under Manage models, the model is indexed, and it must be explicitly synced before it becomes available; other users request access to a model and an admin enables it. Just Ask itself has to be enabled by a system administrator.

A point to check in your tenant: sources differ on data connectivity. An SAP Learning page states that, for live data, Just Ask supports live connections except SAP HANA Cloud and SAP Datasphere, while SAP's newer reference architecture describes analytical insights drawing on data from SAP Datasphere or SAP HANA Cloud. This is evolving with each release — confirm the current support in your tenant's documentation before you plan around it.

What you need to enable it

• A Joule Base entitlement. SAP states that Joule, Analytical Insights is included with Joule Base at no additional cost, and markets Joule Base as included with SAP cloud subscriptions. Contracting still involves choosing an SAP BTP global account (SAP provides one if you have none), so confirm the commercial terms for your landscape.
• The SAP BTP side. Per SAP's reference architecture, Joule runs as a centrally managed instance on SAP BTP, with SAP Cloud Identity Services as the identity broker for both Joule and the SAC tenant, plus a destination and SAML trust for principal propagation.
• Just Ask configured. Once enabled, SAP says the capability is available to any user with an SAC account, and administrators control which models and data Joule can access through the Just Ask configuration.

What it doesn't change

Joule is a new way to ask, not a new way to model. It answers from the models you already have, with the access rules you already set. Two consequences are easy to underestimate:

• Answer quality is model quality. One consultancy write-up recalls the early Just Ask returning "Unknown Entity: month" to a question as simple as "What's revenue by month?". Natural language only works when the model's dimensions, measures and time structure are clear.
• Planning logic stays yours. SAP's published SAC material centres on querying, help, story tasks and scripting. Reporting from SAP Connect in October 2026 focused on Joule Work and line-of-business assistants, with no specific SAC or planning announcement — your versions, data actions and allocations still need to be designed properly.

Make a model Joule-ready: a practical checklist

These are our recommendations, based on how natural-language querying works on a semantic model — not an SAP certification list.

1. Business names, not technical IDs. "Company code" and "Fiscal period" instead of RBUKRS and FISCYEARPER. If your sources use technical fields, map them first — our Live-Ready checker shows that mapping for common S/4HANA fields.
2. Descriptions on dimensions, measures and members. They're what a question has to match.
3. Correct aggregation. A rate summed over months gives a confident, wrong answer. See choosing SUM, AVERAGE, LAST or COUNT.
4. A clean Version dimension. "Forecast vs budget" only works if Actual, Budget and Forecast are modelled as versions — see Actual vs Budget vs Forecast in SAC Planning.
5. A consistent time dimension with the granularity people actually ask about (month, quarter, year).
6. Data access control first. Joule respects the access you configure, so set it before you open a model to natural-language questions.

Where to start

If you're building a model that people will eventually question in natural language, start from a structure with business names, correct aggregations and a Version dimension already in place. Our 64 free SAC templates follow those rules, and the interactive Rolling Forecast demo shows what such a model produces. QRC4 2026 lands on 14–15 November; if it brings Joule changes to SAC, they'll appear in our QRC4 Impact Check and QRC4 hub once SAP confirms them.

Sources (October 2026): SAP Architecture Center — reference architecture for Joule with SAP Analytics Cloud; SAP product pages for Joule, Analytical Insights and Joule Base; SAP Learning — SAP Analytics Cloud AI functionality; ZPARTNER, "Search to Insight vs Just Ask" (2024); Mindset Consulting, "SAC in BDC" (2025); ERP Today report on SAP Connect 2026. We have not tested Joule hands-on. Capabilities, licensing and connectivity change with each release — verify with SAP. Independent project, not affiliated with SAP SE. "SAP", "SAP Analytics Cloud" and "Joule" are trademarks of SAP SE.