SAC AI features explained: Joule, Just Ask, and Smart Predict
- AI, SAP Analytics Cloud
- SAP Business AI
- 10 min read
Philipp Simon
This wiki explains how to use Smart Predict to create automated forecasting models for planning and forecasting in SAP Analytics Cloud (SAC). Our expert Philipp shows how Smart Insights analyzes statistical anomalies in SAC stories at the touch of a button and makes them easy to understand. Another focus of the wiki is on voice-based data queries via Just Ask and the integration of the new AI assistant Joule into the SAP ecosystem. Philipp also highlights the AI-assisted features. Generative AI is used to make tasks such as scripting and commenting more efficient.
Table of contents
- 1. Navigation aid in the AI jungle: Overview of current AI and ML features in SAP Analytics Cloud
- 2. Smart Predict (forecast scenario)
- 3. Smart Insights: What is it and what can it do?
- 4. Just ask: What is it and what can it do?
- 5. Joule: What is it and what can it do?
- 6. AI-powered features
- 7. Conclusion: Evolution rather than revolution through data quality
Monthly financial statements, planning cycles, and operational tasks are the focus of daily work. The applications used are often limited to a familiar set of reports, data, and functions. At the same time, information about new features is pouring in through a wide variety of channels—especially since the increased focus on AI-supported technologies.
SAP Analytics Cloud is no exception. Functions are constantly being expanded, revised, merged, or—if technologically obsolete—discontinued. Solutions such as Just Ask, Smart Insights, and AI-supported functions have created a versatile portfolio, but one that is not always easy to classify in everyday use.
The main challenge is to maintain an overview:
What functions currently exist? What added value do they offer? What limitations need to be considered? And what developments are on the horizon?
This article provides a concise guide and summarizes the most important AI and ML functionalities of the SAC in an understandable way.
The term "artificial intelligence" (AI) describes systems that can interpret information in context and use it to make decisions or take actions. Machine learning (ML) is a subset of AI. It uses past data to recognize patterns and make predictions or recommendations based on them.
First, there are two features that (still) come up repeatedly in conversations or forums, but which have been canceled and patched out of the tenants for some time now: Smart Discovery and Search to Insight. We will therefore not discuss these two technologies further.
We are concerned with the currently available functions.
2. Smart Predict (forecast scenario)
Smart Predict is a feature of SAP Analytics Cloud that is based on automated machine learning and automatically creates suitable models for predictions. The solution independently selects the appropriate method and configuration to efficiently answer business questions. This is based on forecast scenarios, which serve as workspaces. In these scenarios, multiple forecast models can be created and compared for a given question, and the best model can be selected. subsequently Save as a dataset. Each model creates its own visualizations and metrics to understand the result and accuracy. This allows you to create different models within a scenario and save the one with the best results as a dataset.
Depending on the scenario type, imported data sets, live data sets (e.g., on SAP HANA), and, in the case of time series forecasting, planning models can also be used as data sources, provided that the relevant technical requirements are met. There are restrictions on the use of live data sets and planning models in particular.
Smart Predict supports three scenarios:
- Classification: Suitable for predicting categorical events, such as whether a customer will leave.
- Regression: Used to estimate numerical values, for example, to forecast expected sales under certain conditions.
- Time series forecasting: Analyzes historical data and trends to predict future developments over time, such as daily demand for a product over the coming months.
The generated data sets can be used directly in stories, integrated into SAC models, or executed as steps in multi-actions. In this step, it is possible to train predictive models or apply them. Implementation within a multi-action opens up a wide range of different application possibilities.
In planning models, time series forecasts can also be written back into the planning. Smart Predict is included in the SAC standard scope, but is not available in all regions or tenants. Access is granted via the Predictive Content Creator and Predictive Admin roles or via appropriately authorized individual roles.
2.1. When is SmartPredict used?
The use of Smart Predict is particularly recommended in cases where recurring forecasts or reliable pattern recognition from historical data are required. The feature is suitable for prepared, structured questions in the areas of planning, forecasting, or risk assessment.
3. Smart Insights: What is it and what can it do?
Smart Insights is an analysis feature of SAP Analytics Cloud. It uses statistical methods to automatically explain data points. The goal is to reveal anomalies, influencing factors, and correlations that are not immediately apparent in a visualization. The results are displayed directly in the Smart Insights panel within the story and combine written explanations with supporting visualizations.
3.1. Functionality and area of application
Smart Insights is available in Stories for most chart types, table cells, and variances. Imported models, live data sources (SAP HANA), as well as Datasphere and Seamless Planning models in the Optimized Story Experience are supported.
As already indicated, the panel on the right-hand side of the story forms the core of Smart Insights. It presents up to three different types of information with an explanation in natural language and (if applicable) a corresponding diagram.
Contents of the Smart Insights Panel:
Depending on the selected data point, different types of insights can be displayed.
- Change in data point: If a date dimension is available, Smart Insights detects noticeable temporal deviations. The analysis highlights interesting periods such as month, quarter, or year and shows their variance compared to the reference period.
- Key drivers/top contributors: The "Top Contributors" system identifies the most important influencing factors for a data point. It displays the relevant dimensions and the most notable members (top 5 dimensions with up to 10 members). They are sorted according to their deviation from the average.
- Calculation of data points: For calculated values, Smart Insights displays the formulas, aggregation types, and distributions (e.g., histograms, min/max/average) used. This makes it possible to understand how a value is derived.
3.2. When is Smart Insights used?
If individual values, charts, or deviations need to be explained quickly and clearly, Smart Insights can help. Content is presented using graphics and natural language, without the user having to delve deep into the underlying data.
In its simplest form, this feature can be understood as the counterpart to the following Just Ask feature, which creates a diagram or table from a query formulated in natural language.
4. Just ask: What is it and what can it do?
The idea of enabling queries in natural language (Natural Language Query – NLQ) originated from the product of the same name, "Askdata," which SAP acquired in 2022. The technology, originally classified more as an intelligent search function than an AI application, quickly raised high expectations for more natural interaction with company data. With the advent of powerful language models, the potential of such approaches became increasingly apparent. SAP further developed the NLQ function, provided it with regular updates, and established it as an integral part of the analytics environment under the name Just Ask.
4.1. But what exactly is Just Ask?
Just Ask is an interface or function in SAP Analytics Cloud that allows you to query and display data from a model using natural language. Queries are formulated in standard business terms, such as "Show sales in 2024 by region." The results appear immediately as tables or charts and can be transferred to stories, exported, or further examined in Data Analyzer.
Just Ask supports imported data models and SAP Datasphere models, making it particularly flexible to use. Examples of typical queries include:
"Revenue in the last quarter," "Sales figures for Germany and France by product," "Compare actual and forecast versions of a key figure for 2024."
Just Ask can be accessed either via the home page or via the light bulb icon in the upper right header.
4.2. Data quality and model optimization
Optimized modeling is crucial for reliable results. This includes clear, unambiguous object names, consistent date dimensions, and the reduction of redundant or technical elements. Especially in NLQ scenarios, well-structured models facilitate the clear assignment of terms (for example, by assigning synonyms) to key figures, dimensions, and hierarchies.
4.3. Special features in SAP Datasphere
When using Datasphere models, the type of integration (live or tunnel) influences the behavior of Just Ask—especially with regard to data storage, high-performance query processing, and support for synonyms and business rules. Tunnel connections are recommended for more complex scenarios, while live access is more suitable for temporary, light analyses.
4.4. When is Just Ask used?
Just Ask is suitable for quick ad hoc queries where users want to access data without model knowledge or using a story. It can be used primarily to efficiently answer simple analytical questions in natural language—provided that the underlying models are clearly structured and have been prepared for use with Just Ask.
Just Ask simplifies access to data by asking questions in natural language and providing immediate answers in the form of charts or tables. The quality of the results depends largely on clearly modeled, well-structured data. Just Ask is included in the SAP Analytics Cloud feature set and does not require an additional license, but it does require a supported cloud tenant and activation by the administration. With Joule, SAP is expanding these capabilities with a conversation-based AI assistant – the next section shows how Joule complements and extends these functions.
5. Joule: What is it and what can it do?
Joule is SAP's generative AI assistant. It extends the SAP Analytics Cloud with dialog-oriented, context-sensitive functionality. It combines natural language with analytical, transactional, and explanatory functions and is not limited to SAC, but is part of the overarching SAP Business AI.
Joule can incorporate data from SAC models, SAP Datasphere, BW/4HANA, and S/4HANA, providing a unified access point for a wide variety of queries.
(Source: SAP Business AI – https://www.sap.com/products/artificial-intelligence.html)
5.1. Distinction between Just Ask and Joule
Just Ask is the natural language query function within SAP Analytics Cloud. It is used to query data models and display them as tables, charts, or key figures. Each question is interpreted individually.
Joule goes far beyond this and is designed as a complete AI assistant that uses context across multiple interactions, explains relationships, and also supports tasks in other SAP systems. The key point here is:
For analytical questions, Joule uses the Just Ask functionality of SAP Analytics Cloud in the background.
This means that Just Ask is embedded within Joule as an analytical NLQ engine, while Joule extends these capabilities with dialogue management, interpretation, and additional system actions.
5.2. Joule in conjunction with BW/4HANA and SAP S/4HANA
Joule can provide both analytical and operational content.
Analytical content is based on the "Analytical Insights" provided by SAC (and thus Just Ask).
Operational content comes from SAP S/4HANA and is made available to Joule via APIs and live system connections, e.g., to display status information or transactional data. Joule also enables direct access to the associated Fiori applications.
Joule thus combines analysis and operational transparency in a single access point.
5.3. Joule's interaction models
SAP describes four key forms of interaction that Joule supports and combines with each other:
Informational – Provision of explanatory content ("Tell me how...")
Transactional – Support for small tasks or navigation ("Do this simple task...")
Analytical – Providing analytical insights via SAC and Datasphere models ("Give me insights...") à Just Ask
Advanced AI – generative functions such as summaries or script support ("Do advanced tasks...")
These four types of interaction form the basis for the versatile use of Joule in analysis, planning, and operational processes.
5.4. Architecture and embedding
In SAP Analytics Cloud, Joule does not (currently) appear as a separate chat window, icon, or "copilot" within the SAC interface for end users. Instead, Joule can be fully integrated via "Just Ask" in the form of the "Analytical Insights" function on the Analytics page and serve as a natural query interface.
When users work with "Just Ask" in SAC, technically speaking, they are communicating with the generative AI layer that also underpins Joule. However, only "Just Ask" is visible to them as a function of SAC. Joule itself is primarily used as a dialog-oriented assistant in other SAP cloud applications (e.g., Success Factors). There, it uses the exact models that are indexed and approved in SAC for "Just Ask" to answer analytical questions.
In summary, Joule will not be available as a separate chat window in SAC in the current or future release cycles (as of December 2025). Within the SAC, users can access the "Analytical" interaction model via the Just Ask functionality, but not the other options offered by Joule's context-based AI chat. Conversely, users of other SAP cloud applications (such as BDC) can access the Analytical function of Just Ask via Joule.
6. AI-powered features
Another way to use AI in SAC is through AI-assisted features. These are a selection of features designed to facilitate text- and script-based tasks in particular, using generative AI. This is probably not known to many users, as it has never been actively promoted and also comes with a number of requirements:
The features are currently only available in a limited number of data centers (https://help.sap.com/docs/SAP_ANALYTICS_CLOUD/00f68c2e08b941f081002fd3691d86a7/8dcc1f1915b241b3a10c8e5b8a76b062.html?locale=en-US)
Secondly, SAP AI units are consumed for use, which naturally requires that AI units are available.
However, if these requirements are met (tenant on supported data center and AI units available), an administrator can request activation of the functions by submitting a ticket to SAP. Once activation for the selected functions has been approved, the corresponding "Generative AI" authorizations can be assigned in SAC.
6.1. These functions are included therein
- AI-Assisted Data Actions: This feature allows scripts to be created and commented on in Advanced Formula steps. Based on a command, a suitable script can be generated automatically or comments can be written on existing code. This speeds up recurring script tasks and improves the traceability of the logic.
An extension has been announced for the Q1 2026 update that uses generative AI to automatically create a textual description of the purpose and logic of the data action in the settings area of Data Actions (https://roadmaps.sap.com/board?PRODUCT=67838200100800006884&FT=AI&range=FIRST-LAST#;INNO=000D3ABE796A1FE0A5948BE616BDFEB2). - AI-Assisted Commenting: AI-Assisted Commenting is a helpful feature that makes it easier to work with comments in stories. It supports both datapoint comments stored directly in table cells and comments in separate comment widgets. The feature allows you to summarize individual comments, entire comment threads, or—in the case of tables—comments from subordinate hierarchy levels.
- AI-Assisted Calculations: With the "AI-Assisted Calculations" function, calculation formulas can be automatically generated in Data Analyzer using natural language input, or existing formulas can be explained in an understandable way. The AI Formula Assistant generates formulas based on a described requirement and, if desired, provides plain text explanations of complex calculations.
- AI-Assisted Chart Summary: The PowerPoint add-in for SAP Analytics Cloud enables the automatic creation of a comprehensible summary for each embedded SAC chart. The generated text description is inserted as an editable comment and can be recreated when data changes to keep presentations up to date.
6.2. When are AI-assisted features used?
They are an efficient solution for users who want to optimize their analysis, planning, and scripting workflows without having to delve deeply into technical syntax or script logic. Business users, controllers, story designers, and analysts in particular benefit from this feature, as they can generate explanations, formulas, comments, or summaries directly in their workflow. AI provides particular support in situations where routine tasks need to be accelerated, complex content needs to be made understandable, or new artifacts (formulas, scripts, comments, text summaries) need to be created quickly and reliably—without having to leave the application or work context.
7. Conclusion: Evolution rather than revolution through data quality
A look at the current AI and ML functionalities of SAP Analytics Cloud reveals a multifaceted picture: the tools have evolved from pure analysis aids to genuine productivity boosters. While Smart Predict makes complex statistical forecasts accessible to business users, Just Ask and Joule's Analytical Insights massively lower the barrier to spontaneous data access. This portfolio is complemented by AI-assisted features that save valuable time, especially when creating content, whether it be scripting or commenting.
But despite all the enthusiasm for generative AI and automation, one old truism of data processing remains true: AI is only as good as the database it accesses. Features such as Just Ask in particular clearly show that clean modeling and clear semantics are the basic prerequisites for reliable answers.
For companies, this means that the technology is ready. The key to success now lies in creating the organizational and technical prerequisites (such as AI units and data quality) to establish these tools not just as a technical playground, but as an integral part of modern planning and analysis.
Want to learn more about AI Units & Co. and find out how to successfully launch your AI project?
Then book an appointment with our SAP AI expert Robert Kehrli here.
