Unleashing AI on Live Data: The Power and Limits in SAP Analytics Cloud

Unleashing AI on Live Data: The Power and Limits in SAP Analytics Cloud

SAP Analytics Cloud (SAC) empowers users with powerful AI and machine learning capabilities, enabling deeper insights into their data. While imported data models traditionally offered the most robust AI experience, SAC has evolved to integrate AI with live connection data, albeit with certain limitations.

Harnessing AI with Live Connections:

Live connections in SAC provide real-time access to source data, ensuring your analyses are always up-to-date. But can you leverage AI on this dynamic data stream? The answer is a qualified "yes."

Here's a look at the AI features available for live connections:

1. Smart Insights:
   • This feature acts as your automated data detective, diligently scanning visualizations for patterns, trends, and outliers.
   • It operates on the data already loaded into the story, providing quick, insightful observations.
   • Furthermore, Smart Insights offers automatic chart recommendations, streamlining the visualization process.
2. Smart Discovery (Limited):
   • While not as comprehensive as with imported data, Smart Discovery offers basic automated analysis on live data.
   • Users can conduct simplified key influencer analysis and detect basic correlations, providing a glimpse into the underlying data relationships.
3. Search-to-Insight:
   • This intuitive feature allows you to explore live connection data using natural language queries.
   • Simply ask questions about your data in plain language, and SAC will generate relevant visualizations and insights.

Navigating the Limitations:

While these features offer valuable insights, it's crucial to understand the limitations of applying AI to live connections:

1. Processing Constraints:
   • AI operations are confined to the data already retrieved into the story.
   • Analyzing the entire source dataset is impossible if it exceeds the loaded data.
   • Consequently, performance may be slower compared to imported data models.
2. Feature Restrictions:
   • Advanced predictive scenarios, such as time series forecasting and complex classification/regression models, are generally not supported with live connections.
   • These advanced features typically require the more stable environment of an imported data model.
3. Data Blending Limitations:
   • AI features cannot operate across multiple live connections or on blended datasets from different sources. This restricts the scope of analysis when dealing with diverse data landscapes.
4. Recommendation:
   • For advanced AI/ML use cases, consider creating imported models or utilizing SAP HANA PAL (Predictive Analysis Library) at the source.
   • Data snapshots can also be a valuable tool for comprehensive predictive analysis.

The Optimal Approach:

The most effective strategy often involves a hybrid approach:

• Utilize live connections for real-time operational analytics, ensuring up-to-the-minute insights.
• Create targeted imported models for specific advanced AI analysis needs, enabling complex predictive scenarios.

By understanding the capabilities and limitations of AI on live connections, you can leverage SAC's powerful features to gain deeper insights and make data-driven decisions.