In this installment of our blog series about SAP CoPilot, we will explore in more detail the topic of natural language processing (NLP). The goal of NLP is to enable users to communicate with software in an intuitive way, that is to say, in languages that people use to talk to each other instead of in special computer languages.
Getting stuff done, today and tomorrow
How do you currently apply for a day of vacation in your company? Personally, I go into our corporate portal, click on “Employee Services,” there I see an SAP Fiori tile called “Time & Vacation.” Inside that tile I click a link called “Leave Request.” Then I see an overview of a dozen or so of my past leave requests as well as a button called “Create New Leave Request.” I click on that and then a new screen appears with a calendar view of my teammates’ planned vacation. At the bottom of the screen, I can specify the exact days and the type of leave request that I want.
Now don’t get me wrong. This is a very good user interface! It works really well, provides me with relevant information for the task at hand, looks good, and is easy to use. But let’s bring natural language processing into the experience and see if it could be better…
As an optional component of Fiori 2.0, SAP will introduce CoPilot, the integrated digital assistant for the live enterprise. So let’s imagine applying for vacation with SAP CoPilot:
You: Click on the SAP CoPilot icon in SAP Fiori and say, “Hey, I need a day of vacation!”
SAP CoPilot: “When would you like to take a day of vacation?”
You: “Next Friday.”
SAP CoPilot: “Do you want me to create a leave request for you for Friday, November 25th?”
SAP CoPilot: I have created the following leave request for you and sent it to your manager.”
You: “Oh. Can you also tell me how many people in my team are also taking next Friday off and how many days of vacation I have left this year?”
SAP CoPilot: “Two of your teammates have scheduled vacation for November 25th. You have 10 days of vacation left this year.”
You: “OK. Thanks.”
SAP CoPilot: “You are welcome.”
So what is going on here? First of all, you can interact with the system more or less as if it were a real person. By applying state-of-the-art natural language processing and understanding techniques, SAP CoPilot would know that when you say, “a day of vacation,” your leave request is not for a sick day, but for a day off. It knows that “next Friday” is “November 25th.” You also don’t need to scan through tables of data you may or may not be interested in. SAP CoPilot immediately delivers the information that you requested. You can speak to SAP CoPilot in a way that is natural and intuitive, and you don’t need to navigate from one screen to another to access the leave request app. SAP CoPilot goes directly to the right place. Using a digital assistant dramatically unifies and simplifies the user experience.
The scenario above exemplifies a question and answer interaction between a human being talking in a natural way with an intelligent bit of software that understands the user’s business context and has access to the enterprise system. Asking questions and getting useful answers that help you to quickly reach your goal is a great step forward, but how can this get even better? Let’s project this chat scenario a bit farther into the future, where the SAP CoPilot becomes a true virtual assistant.
Beyond Questions and Answers
According to Venessa Micelli-Schmidt, Product Expert for SAP CoPilot, “Based on your previous actions, SAP CoPilot could proactively provide information or recommendations that relate to what you are currently doing or that are triggered by a specific point in time.” So, getting back to our leave request example, SAP CoPilot could learn that you usually ask about your colleagues’ planned absences when you apply for your vacation. It could provide that information to you each time you request for time off without you having to ask it to do so. Want it to stop? Just say so.
As SAP CoPilot is aware of the user’s business context, it could offer relevant insights just when you need them based on your role, context and current business situation. Micelli-Schmidt predicts that “State-of-the-art machine learning techniques will make SAP CoPilot smarter over time and allow it to identify situations in your working environment where you might need assistance.” That means that digital assistance has the potential to go far beyond simple scenarios such as leave requests. Imagine one of your contracts is about to expire. SAP CoPilot could, for example, not only notify you, but also recommend that you have a look at the contract and prolong it.
Have a digital assistant reach deep into business tasks, especially complicated and business-critical ones, may sound scary. But if the situation or question is unclear, SAP CoPilot will answer back with a question to clarify. And this interaction will incrementally make SAP CoPilot more competent, ultimately helping you to do your job better and faster than ever before.
You can also stay on top of what SAP is doing with regard to user experience and design by following @SAP_designs on Twitter.