Product Information
SAP Smart Predict – Recent Deliveries, Outlook Webcast Summary
This was an ASUG webcast last week with Jan Fetzer , Product Management, SAC (SAP Analytics Cloud), focusing on SAC Predictive. Thanks to SAP’s Ingo Hilgefort for arranging this.
Review of where SAP stands with Smart Predict, with an outlook, new developments
Figure 1 Source SAP
Augmented Analytics – Overview & Positioning
Visual – business led, diagnostic analytics, interactive, focuses on best visualization of data
Tools serve self-service needs
Structured data
Adoption rose
35% adoption
Advent of big data, unstructured data, structured data, deluge of data, focus of analytics changed to be machine-led
Pervasive, diagnostic analytics
Significant insights for users
Patterns, mining data automatically
New interaction schema for interacting with system
Systems can suggest, user context, embedded in apps
See higher adoption
Figure 5 Source SAP
Business led patterns, on the left
Added additional learning, intelligence, system can suggest, explain data points, relevant insights
On the verge of alerting people about changes in data
Complete predictive engine at the heart of SACFigure 6 Source SAP
See ways people use the system, seeking the guidance from the smart capabilities within SAC to find relevant insights on data
Figure 7 Source SAP
3 buckets of functionality shown above
Converse with system, search to insight – on desktop or mobile, type question and system understands what you mean and generate best representation
Activate smart insight on chart showing textual representation on screen
Investing in this to make it smoother, custom vocabulary
Automated insight – uses machine learning, AI, to find explanations to see what drives data point
Predictive – build a predictive engine in the heart of SAC – look forward instead of traditional BI solutions
Can do this at scale
Predict what will happen in the future
Smart Predict Today
Figure 8 Source SAP
Smart Predict, forecast time series
Classification – classify customers, products, which customers will buy – yes/no
Regression – continuous signal – how will a number be at a certain point in time
Predictions in a dataset can be copied to planning models – tool to do that are called data actions; will be simpler in Q3 release
Figure 9 Source SAP
Problem focus of business outcome – e.g. anticipating customer behavior
Figure 10 Source SAP
Classification – yes no decisions, on off decisions
Regression – numeric values, influencers of those numbers
Time series – generated as a multiple time series – by business area, product group, etc.
Data this can be built upon is shown above; can connect directly to Live HANA system without moving the data to the cloud
Train the model
Smart Predict Roadmap
Figure 12 Source SAP
Where is SAP investing
Will soon release integrate with planning workflows, already with beta customers; planned for August release
Planning – what should happen based on company targets, while monitoring that, forecasting shows do current developments compare
Figure 15 Source SAP
Scenarios seeing with customers
Figure 16 Source SAP
Can use Smart Predict with planning models
Business focus
Beta closes in 2 weeks
GA planned for Q3
See predict scenarios in file system
Refreshing of data sets, live datasets (BW/4HANA + calc view support)
What do you think?
Thanks Tammy. Kind regards, H