Infosys’ blog on industry solutions, trends, business process transformation and global implementation in Oracle.

« Artificial intelligence - the theme for next generation CRM | Main | Weblogic Password Re-set or Informatica Connect String change, we have a solution for you!!! »

Einstein Data Discovery - An impeccable "Data Scientist" from Salesforce


With an ambition to bring Artificial Intelligence to business users, Salesforce recently introduced the Data Discovery as Einstein Analytics product which empowers business users to spontaneously discover unidentified prospects and have richer vision by getting down to the bottom of organization's huge volumes of data instantly. Before I get down to the nitty-gritty about Einstein's Data Discovery, I think a simple definition of Artificial Intelligence is worth to mention here.

"Artificial Intelligence is the widespread concept of machines modeled in imitation of human mind or the way human mind accomplish tasks and considered to demonstrate intelligent behavior in a smarter and faster way"

Realistically human brain is not casted to analyze and match with the speed which the data is getting big every day across all the domains be it retail industry, manufacturers , financial bodies or health care.

Despite the fact that AI is upright at dissecting data and creating insights, business is not able to gain benefits from same. It is extremely challenging for business users to understand facts, scenarios and insights the way they are presented by AI. Since long analysts and data scientists are plugging this gap. To crack the concern of limited skilled resources, availability of data scientists and empower every business user to apply intelligence in parallel with the speed of business, now we have the perfect data scientist which is officially known as "Salesforce Einstein Data Discovery".

There are enormous "W" and "H" questions business executive's wants to know or rather they should know to make more mindful business decisions and they need to understand the questions in a narrative way. "Einstein Data Discovery is one narrative tool which allows them to discover insight and analyze massive volume datasets in minutes without getting troubled in building statistical data models".

It handles descriptive, diagnostic, predictive and even the prescriptive analytics which makes it capable of answering why something happened, what could happen in future and what action should be taken? It not only has easy guided recommendations for getting deep insight and answers but with Einstein Data Discovery one can even get to know the out of sight questions too.

I remember an instance from the conversation I had some time back while working closely with the VP of operations of a large beverage company. He was supposed to answer company's CFO in a very short time the reasons for the falling margins and what are the steps required to recover. The discussion was around how he made his team of data analysts and business executives to work overtime and burdened them with heaps of data and millions of transactions to segregate the potential problems causing a slip in margin. After spending days and lots of brainstorming from team members, he wasn't able to get clear answer on what actions should be taken even though he and his team known the business best but still was not able to leverage the domain knowledge.

In all such any many more daily business scenarios, having Salesforce Einstein Data Discovery readily available make it so relevant that you only need to load your data to Einstein Discovery and specify the variable you are looking for. By the time you get yourself a cup of coffee, Einstein Discovery is ready with the story line.

It unlocks the appropriate metric and its influence on business in an organized manner with a logical flow packed with insight about organization's data.  It points you to correct direction, provide relevant pattern and recommend answers and let you leverage the tool to take decisions based on what is statistically appropriate.

Einstein Data Discovery.JPG


Storytelling Behavior:  One of the key feature which makes Einstein discovery a bridge between business users and AI is the narrative explanation of insights which is what every business user needs to be self-sufficient and less dependent on data scientists. Once you are done importing dataset to Einstein Discovery, it assesses your data from every possible angle, do statistical test, detects patterns and evaluate the improvement opportunities. Once the analysis is ready which in terms of data discovery is called as a "story" is presented to you. It does not merely show the data but highlight the importance of it. The visualization includes a graphical representation of facts for example a bar chart with a descriptive text which is easily understandable to business users.

Automatic Discovery: Another significant ability is to automatically scan the data for quality problems, analyze the millions of datasets quickly and provide the underlying cause behind business outcome, what is next in line and what are the recommended actions. There is a recommendation section which gives insights on What Happened? What Changed over the time? Why it happened, what could happen and how can I improve it? It gives you the option to relate these questions to different variables to slice and dice the facts to dig down deeper and getting a broader picture

Flexibility to get data from any source: One will never have to worry about the source of data when working with Einstein data Discovery. It can take data direct from Salesforce or importing from a CSV file. It also works with any structured database viz Hadoop or SAP.

Natural Language Support: All the generated reports and insights are explained in natural language and can be exported in PowerPoint presentations and can be shared effortlessly.

Smart Data Quality Check: When you pull the data to get the stories created, Einstein Discovery automatically scan the data for data quality problems to see if all rows and columns are what we need. It identifies ways to improve the quality of your data-set and recommend the cleanup actions. For example while loading customer data, if the "Age" column has some text data which we might consider changing to number for analysis.

User Feedback Loop: Einstein Data Discovery have this great feature to refine the data model based on the feedback provided by users. Based on experience and expertise you can give a thumbs up or down to get more accurate analysis.

Salesforce Native: Einstein Data Discovery has native connectivity with Salesforce which allow users to work in parallel on Einstein discovery to get the insights and quick navigation to Einstein Analytics for taking required actions.


Einstein Data Discovery is certainly my personal data scientist now! I tried to provide as much information I could post here and will be happy to answer if there are any advance questions around this topic. Please write in comments section.

Comments

Impressive, and useful as well.

Very well explained Saryu.Is there any detailed post on Einstein Analytics too

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

Subscribe to this blog's feed

Follow us on

Blogger Profiles