Microsoft Power Virtual Agent Demo
Starting the Socrates Digital™ Demo in Teams
This is a demo of a Socrates Digital™ application that was created in Microsoft Power Virtual Agent and deployed in Microsoft Teams. This is what is called a “no code” implementation of a Socrates Digital™ application. There is some logic that needs to be configured but no code in a general-purpose programming language such as Java, C, or VBA is used. If you have Microsoft Office 365, then this application is essentially free since a version of Microsoft Power Virtual Agent comes with Microsoft Teams.
This demo begins with Socrates Digital™ asking the user what it can help with today. The knowledge base for this demo is about evaluating investments in the commercial real estate industry. The user responds with a request for help with investments.
Deciding What a Good Solution Looks Like
In this screen capture, Socrates Digital™ asks the user how the decision to invest going to be made. The user responds by saying the investment decision will be made by determining if the investment will boost total revenue enough to pay back the investors. Socrates Digital™ then asks precisely how this statement relates to the question at hand – in other words, it asks the user to state all the decision criteria. The user responds by saying it must not only boost revenue enough to pay back the investors, but it must be enough to meet the projected rate of return for the investors.
Picking the Socratic Questions to Answer
In this screen capture, the user tells Socrates Digital™ that the investment under consideration is the Rosebud Motel chain. Socrates Digital™ responds by asking what the user would like to work on next. The options for the user to select are from the Socratic categories of questions that relate to the information that is available for the Rosebud Motel chain1, the assumptions that are known, the concepts that are applied, the conclusions that are developed, the implications of those conclusions that are predicted, and the viewpoint that is made from answering the questions in all these categories.
1Any similarity to the Rosebud Motel in the situation comedy “Schitt's Creek" (TV Series 2015–2020) is merely coincidental.
Finding Facts that Support Concepts
This screen capture shows that the user selected to answer questions in the information category. Socrates Digital™ follows up by asking what information the user has. The user responds by saying that information about occupancy rates and revenue are available in a spreadsheet for the Rosebud Motel chain. Socrates Digital™ then asks for some specific data from the spreadsheet that will provide some insight into occupancy rates and revenue for the Rosebud Motel chain.
Identifying Assumptions for Analyzing Information
In this screen capture, the user responds by noting that one of the motels in the Rosebud Motel chain has low occupancy rates which seem to be related to low revenue rates. Now, Socrates Digital™ switches gears and lists all the assumptions it knows about that relate to evaluating investments in the commercial real estate industry. For this demo, the listed assumptions were supplied by the analysts that created the knowledge base in this area for Socrates Digital™. The user is now asked if any of these supplied assumptions can be used to make sense of the data point suggesting that low occupancy rates seem to be related to low revenue rates. The user responds that the assumption “Occupancy rate is positively correlated with total revenue” does explain the data.
Checking if Assumption Holds for Larger Dataset
Socrates Digital, in this screen capture, asks the user a question about applying this assumption. Socrates Digital™ uses this answer by the user to later calculate the level of confidence the user has in this assumption for evaluating if the Rosebud Motel chain is a good investment. In a follow up question, Socrates Digital™ now asks the user if this assumption holds for a larger dataset. The user responds that the occupancy rate is positively correlated with total revenue for a larger dataset.
Developing a Conclusion and Evaluating the Logic
This screen capture shows that Socrates Digital™ now moves toward getting the user to state a conclusion based on the analysis of the information, assumptions, and concepts (which was skipped in this demo) gathered so far. Socrates Digital™ will continue to work with the user to analyze additional information, assumptions, and concepts until the user has no more additional ones to add. Also, skipped for this demo, were the questions put to the user about the implications that can be predicted from the conclusions developed.
After the user states a conclusion, Socrates Digital™ asks the user to evaluate the logic of the conclusion. Socrates Digital™ will also use this answer later to calculate the level of confidence the user has in this conclusion for evaluating if the Rosebud Motel chain is a good investment.
Restate Implications to Create a Viewpoint
After making sure that there are no more information, assumptions, concepts, conclusions, and implications to consider, Socrates Digital™ asks the user to restate the implications of the developed conclusions into an overarching viewpoint that summarizes the results from all the questions that Socrates Digital™ has asked the user. Socrates Digital™ now goes to work to calculate a value for each Socratic questioning category -- information, assumptions, concepts, conclusions, and implications.
Calculating Overall Confidence in Viewpoint
In this screen capture, Socrates Digital™ is calculating a value for each Socratic questioning category -- information, assumptions, concepts, conclusions, and implications – and a final overall level of confidence that the user has in the viewpoint for deciding to invest in the Rosebud Motel chain. In this demo, the overall level of confidence is .84 – a high level of confidence that this viewpoint is a good solution. In the longer demo of this example, the examination of the average daily rates and the revenue by available room are examined. Also in the longer demo, other users have built on this viewpoint with more information, assumptions, concepts, conclusions, and implications.
Deciding if Question at Hand is Answered
This screen capture shows Socrates Digital™ asking the user if the overall confidence in this viewpoint answers the question at hand – deciding whether to invest in the Rosebud Motel chain. Not shown in this demo is the discussion about the implications of this decision. As this point, Socrates Digital™ has brought the users to a solution for their problem and has stored all the results of the problem-solving process in SharePoint and all the calculations in Excel. The results can now be accessed for review and reuse – providing the basis for organizational learning.
Storing Session Results in SharePoint
This screen capture shows that this implementation of a Socrates Digital™ application in Microsoft Power Virtual Agent also uses SharePoint to store the results from the Socratic questioning sessions with the user.
Managing Calculations in Excel
In this screen capture, we see that the calculations by Socrates Digital™ are done with Microsoft Excel. We also see that the weights for each factor – or category – can be set and empirically adjusted in Excel. This is true of the multiplier factors as well. This gives the developers of Socrates Digital™ an easy way to “fine tune” their application of Socrates Digital™ to fit the problem area that they are working in.
Developing in Microsoft's Power Virtual Agent
This screen capture shows the development environment for Microsoft’s Power Virtual Agent. If compared to the first screen capture above, we see that it asks the user what it can help with today. Next, the logic shows that it offers the options of bonds, investments, and stocks to the user. These options are branches of a tree that are created through a “click and drag” environment that is called the Canvas in Power Virtual Agent. With this environment, the developer can create a visual representation of the logic that presents user questions, captures user responses, and presents follow-on questions based on those user responses.