AI is profoundly transforming the way we work, analyse and process data. Fortunately, SAP did not wait for its partnership with Mistral AI to integrate the power of AI into SAC and, more broadly, into all of its solutions.
For many years now, SAP Analytics Cloud has enabled companies to automate their data analysis, anticipate trends and improve their performance.
Mathieu Le Berre, Consulting Director at Sileron and Business Intelligence expert, has broken down the various AI features available in SAC and provides concrete use cases to help you grow your business.
The role of AI in SAC
SAP Analytics Cloud has always offered an innovative and, above all, scalable decision-making solution based on AI and machine learning.
By connecting to key enterprise data sources such as ERP and/or external databases, SAC provides a comprehensive, real-time view of business performance indicators.
Ultimately, implementing SAC means choosing to leverage your company’s data, integrating AI in a concrete and operational way, and strengthening your long-term competitiveness. Because today, it is quite clear that AI is no longer an option for companies, but rather a strategic lever.
To achieve this, SAC relies on several robust features that are directly embedded in the solution:
- Smart Predict
- Smart Insights
- Just Ask
- Compass
Smart Predict in SAC, predictive AI
SAC’s Smart Predict feature enables companies to use their historical data to create statistical models for predicting the future.
Smart Predict uses machine learning to function. This means that, at regular intervals, it will be possible to train the system so that it gradually understands the enrichment of data and improves its statistical results.
Once this work is complete, the Smart Predict feature is able to:
- edit a statistical model
- explain this statistical model in a dedicated screen
- provide forecasts over a defined period
- indicate a confidence interval for the future based on past data
Possible applications through the three main approaches of Smart Predict
=> Approach No. 1 in Smart Predict: classification
In Business Intelligence, a classification scenario allows you to predict a category or class, i.e. to answer a closed question such as “Yes/No” or “Class A, B, C”.
Examples of real-life cases:
| Area | Question | Classification type |
|---|---|---|
| HR | Will an employee leave the company? | Yes/No (binary) |
| Marketing | Will this customer respond to my campaign? | Yes/No |
| Finance | Is this file risky? | Low/Medium/High |
| Maintenance | Will this machine break down in 30 days? | Yes/No |
How does classification work in Smart Predict?
1- We start with a history
We start with past cases where we know both the characteristics and the outcome (e.g. employees and whether or not they left the company).
2- The SAC engine analyses patterns
It looks for what differentiates “Yes” cases from “No” cases, as well as factors that increase the likelihood of a positive outcome.
3- We then apply the model to new cases
For each new situation, SAC will then provide the predicted class (Yes/No) and the associated probability (e.g. 87% chance of it being “Yes”).
=> Approach No. 2 in Smart Predict: the regression scenario
In SAC, a regression scenario allows you to predict a continuous numerical value based on known data. Unlike classification (Yes/No), here we are looking to estimate a figure (amount, score, duration, etc.).
Examples of real-life cases:
| Area | Question | Prediction type |
|---|---|---|
| HR | What will be the expected salary of an employee? | Value (€) |
| Sales | What will be the average value of an order? | Value (€) |
| Projects | How many days behind schedule will this project be? | Duration (days) |
| Finance | What will be the annual training cost per department? | Value (€) |
How does the regression scenario work in Smart Predict?
1- We start with a history
We have a table where the target value is known (e.g. cost, salary) with explanatory columns (age, experience, position, etc.).
2- The SAC engine learns to make predictions
Smart Predict identifies the relationships between known characteristics and the target value to be predicted. It detects trends, correlations and cross-effects.
3- The model is then applied to new cases
For each row, SAC provides a predicted value. For example, “this employee will probably cost £3,700 in training”.
=> Approach No. 3 in Smart Predict: Time Series Forecast
Another really interesting approach in BI, the Time Series Forecast scenario (also known as “time series forecasting”) allows you to predict a numerical value over time, based on its chronological history.
Real-life example:
| Area | Question | Type of prediction |
|---|---|---|
| Sales | How will monthly turnover evolve? | Amount per month |
| HR | How many new hires can be expected in the next quarter? | Volume |
| Logistics | How many parcels will be shipped per day? | Quantity |
How does Time Series Forecast work in Smart Predict?
1- The SAC engine analyses patterns in actual data
Smart Predict detects upward/downward trends, seasonality (months, days, etc.) and cycles with breaks or peaks.
2- SAC applies an identified statistical model to project future data on each data set
SAC delivers expected values for future periods associated with a confidence interval.
Smart Insights in SAC, explanatory AI
Smart Insights is an augmented intelligence feature in SAC that automatically explores the explanatory factors of a phenomenon, discovers hidden relationships between data, and generates ready-to-use visualisations and insights.
Again, thanks to machine learning, Smart Insights automatically analyses the underlying data and identifies the factors that have an impact.
Possible applications of Smart Insights
Examples of real-world cases:
| Area | Question asked | What Smart Insights provides |
|---|---|---|
| Finance | Why did costs skyrocket in March? | Main contributors (e.g. transport costs + Europe region) |
| HR | Why is absenteeism higher this week? | Dominant segment (e.g. warehouse A + night shift employees) |
| Sales | Why did the margin drop on this product? | Customer, channel or area contributing negatively |
How does Smart Insights work?
1- The user clicks on a value in a graph or table and then selects “Smart Insights”.
2- SAC automatically analyses the available dimensions
It also identifies the main contributors to the selected value.
3- SAC displays the results in a side window and presents calculations
The calculations may be of the type: “Which segment best explains this value?” or “What distinguishes this value from the rest?” The engine then performs a contextual statistical analysis without the need to create a model or script.
Just Ask in SAC, the conversational analytics AI
Just Ask is the conversational analytics feature in SAP Analytics Cloud that offers the ability to query in natural language. It allows users to freely ask questions about their data, as if they were talking to an intelligent assistant.
Thanks to natural language processing (NLP) and AI in SAC, Just Ask is able to interpret query intentions, understand the business context, and automatically generate appropriate visualisations such as tables or graphs.
This feature simplifies access to BI, allowing non-data experts to obtain reliable, contextualised information.
Practical applications of Just Ask
With Just Ask, any user with access can quickly obtain visualisations or key values from a question asked in natural language.
Examples of practical applications:
| Area | Question asked in Just Ask | Result |
|---|---|---|
| Sales | ‘Top 5 customers by revenue in 2024’ | Chart of top customers |
| Finance | ‘Show net margin by quarter’ | Timeline |
| HR | ‘Number of employees by department’ | Table or bar chart |
How does Just Ask work?
1- The user opens the Just Ask search bar
They ask a question in natural language (English, French, etc.)
2- SAC analyses the syntax of the question
It identifies the dimensions (e.g. country, date) and measures (e.g. revenue, margin) to be used to answer the question and creates a dynamic analytical query.
3- SAC provides a response that matches the question asked
SAC generates a numerical result (KPI) or an automatic visualisation (bar chart, pie chart or curve).
SAC generates a numerical result (KPI) or an automatic visualisation (bar chart, pie chart or curve).
Compass in SAC, AI for decision support
Compass is a decision support feature in SAP Analytics Cloud. It guides users in exploring their data by suggesting intelligent analytical recommendations.
Based on the Monte Carlo method, it simulates the impact of uncertainties on KPIs.
Practical applications of Compass
With its user-oriented approach, Compass facilitates the discovery of important data, speeds up analysis and helps teams focus on the most valuable information.
Examples of concrete cases:
| Area | Example question addressed with Compass |
|---|---|
| Finance | What is the probability that our ROI will exceed 12%? |
| HR | What is the risk of not reaching the planned headcount? |
| Sales | What is the impact of volume and price fluctuations on our turnover? |
| Projects | What is the probability that the project will exceed 9 months? |
| Supply Chain | What level of stock should we plan for to cover variations in demand? |
How does Compass work?
1- The user defines a target KPI
(e.g. net profit)
2- They configure several uncertain factors called “drivers”
Drivers are variables whose values can fluctuate, such as price, volume, cost, exchange rate, etc.
For each “driver”, a range of variation (minimum/maximum) and a distribution (uniform, normal, etc.) must be specified.
3- Run a Monte Carlo simulation
Compass generates several thousand random scenarios. This allows it to calculate the KPI value each time based on the drivers and display the complete distribution of results.
To conclude on AI in SAC
Thanks to enhanced analytical features such as Smart Predict, Smart Insights, Just Ask and Compass, SAC democratises access to advanced real-time analysis, automates and improves the quality of decision-making.
According to Mathieu Le Berre: “In concrete terms, AI in SAC does not replace business expertise but amplifies it by facilitating access to readable and predictive data.”
Would you like to learn more about how SAP AI can revolutionise your business processes?
This article was written with the help of Mathieu Le Berre, Consulting Director at Sileron and Business Intelligence expert
Since 2019, Mathieu Le Berre has been Consulting Director at Sileron, where he is the go-to person for Business Intelligence solutions.
An SAP expert since 2008, he is equally skilled in business issues and the technical aspects of cloud solutions.
