Microsoft Ignite 2017: AI, Chatbots, And A Little VR

Conferences like Microsoft Ignite 2017 allow tech professionals of all kinds to come together and share knowledge. From enterprise IT solutions to artificial intelligence (AI) and deep learning breakthroughs, Ignite has provided an excellent opportunity for someone like me, a software and technology researcher and writer, to learn more about the world of chatbots and of neural networks, and even experience a little VR (featured below).

Warning: I may or may not get a little freaked out. Excuse my language.

VR and CPQ

Although I follow AR/VR closely, getting a chance to actually put on a VR headset is an absolute trip. In this case, the demo was set up by a company that has a configure, price, quote (CPQ) solution. Now, you might be wondering: what does a CPQ tool have to do with VR?

In a general sense, CPQ software is used in sales departments to accelerate the sales process by helping determine the needs of a customer and providing them with accurate, custom quotes that best fit those specific needs. While this is already beneficial to both parties, VR can take it one step further by letting customers virtually browse inventory and customize products. Real-time quotes could then potentially be generated, allowing for an even faster quote-to-cash process, and ultimately a better customer experience.

AI Chatbots and Vertical Industries

In a session titled, “Bot Framework Patterns And Practices Straight From Our Customer,” Robert Standefer and Cindy Noteboom spoke on the AI framework and evolution of chatbots in areas like customer service, help desk, and vertical industries such as retail and automotive.

About a quarter of the session was spent going through how to build a chatbot that can use cloud platforms like Azure and AWS to pull data through a CRM, as well as the powerful business insights that can be provided by business intelligence (BI) software. However, the majority of the session discussed the importance of user experience (UX) when determining whether a bot is successful or not.

Using retail as an example, Noteboom discussed the use of a bot named “Cami” who provided customer support for a large electronics company. Aside from a solid framework, Noteboom emphasized the importance of inclusive design, personality and proper branding when creating a bot that is ultimately the face of your customer service department.

Deep Learning Built with Logic

Rafal Lukawiecki believes that the future of data science hinges on our ability to insert human logic into the equation. In his session, “Artificial intelligence: Past, present, but is the future just about machine learning?,” he discussed the origins of AI and ways to combine probabilistic programming languages with neural networks and deep learning.

Lukawiecki talked about the fact that as a business changes, machine learning models evolve, and software then begins to make “intelligent” decisions. As a result, organizations must keep changing models because they become obsolete so quickly –– AI changes the market in which you work, your customer base and your competitors strategies. Although this can be massively beneficial, the problem he sees with deep learning, and AI in general, is the inability to predict how a program will make decisions. This leads to system constantly being retrained after errors inevitably begin to show up.

One example Lukawiecki brings up is the use autonomous cars, where actual human lives are at stake. In this case, if the program running in a self-driving car is not perfect, it will need to be retrained over and over again. The scary part is that every time a neural network needs to be retrained, people are getting in accidents and potentially experiencing injury or death.

However, Lukawiecki feels that he has a solution: a mixture of programming languages that are probabilistic and include human logic and deep learning to do the heavy lifting. His hope is that more companies begin to take a “long-view” approach to creating neural networks that last.

If you were not able to attend Microsoft Ignite 2017, check out the website to get full listings of all the speakers and sessions.

?G2 Crowd is the leading B2B software review site, with more than 230,000 validated user reviews and read by nearly 1 million visitors per month. Learn more about G2 Crowd here.