Product Information
Search to Insight Use Case
Editor’s Note: These features are reflective of Search to Insight as of version 2019.10. For information on previous versions, see our page on augmented analytics.
John Smith is a sales analyst for a global ecommerce clothing company. He was just asked by his manager to prepare key findings for a board meeting taking place later that week. John started a week ago and is not yet familiar with the sales figures of the company. He knew he had to come up with something fast!
Using SAP Analytics Cloud, John turns to its augmented analytics capabilities to make his work faster and efficient.
Using Search to Insight
Search to Insight brings conversational AI to SAP Analytics Cloud, allowing you to create a story by asking questions in natural language. It’s like a search engine, where you can just ask questions about your data.
This feature works by processing your natural language queries and providing you with answers in numeric point or chart form. Search to Insight uses auto-complete suggestions to match words or phrases in your question to measures and dimensions in your data. After interpreting your question, you would see more than one option in the results panel.
John starts his analytics journey by launching Search to Insight. He is able to do this by either clicking on the dedicated lightbulb icon in the tool bar or through the search bar on the home page.
Being new to the product John is not quite sure how to get started, so he clicks on the surfaced recommendation “net revenue for 2019” to discover that the company made $190M this year.
He wants to get a better idea of how sales are distributed so he asks, “What is our net revenue by sales region for 2019.”
Before executing the search, John notices that autocomplete suggestions from other models are appearing.
Knowing that he only wants to pull information from “ProfitandLoss2019,” John locks in the search to the model level. This will narrow down his results and restrict his search to the context of the “ProfitandLoss2019” model.
The resulting bar chart has over 50 regions!
John is only interested in the largest markets generating the most revenue, so he uses the rank and sort feature to display the top regions by adding “top 10” to his search.
John now has a broad idea of the net revenue figures for 2019.
To get to the next level of detail, he would like to see the changes over time. Thus, the next question he is asking is “net revenue by date”.
In the last few months net revenue has increased significantly, this leads him to believe that the company has exceeded their goals. If true, it would be a great finding to bring up during the meeting.
To investigate his theory John needs to compare the actual and projected numbers from Q1. He does this by specifying the version types and time period in his question “actuals vs forecast net revenue for last quarter.”
The provided chart confirms that the company exceeded their projected net revenue.
John is aware that there is more he can ask but does not recall the details of the model. He opens the model details drawer and sees that there are 45 measures and 5 dimensions.
Instead of typing out his next question, he selects the “net revenue” measure token and the “product” dimension token to generate his query.
John is almost prepared for the board meeting but forgot to write down the answer to his first question “net revenue for 2019.” Luckily, his session search history was saved by SAP Analytics Cloud. John can just scroll up to find the result without having to conduct the search again.
Traditionally, users would require BI expertise in order to find the correct data and build the appropriate visualizations.
After solving his challenges with SAP Analytics Cloud, John can quickly discover important trends in his data and accelerate his workflow by asking questions in natural language. With all the insightful findings collected, John is surely to impress at the board meeting!