By day 3, Dreamforce usually starts to wind down. The concerts, the parties, the keynotes, the sessions – it all begins to feel like a cocktail party that will not stop. This year, Dreamforce organizers made some changes in the agenda to address the inevitable ennui that sets in after so much excitement. In years past, the headliner spoke on the first day of the event. At Dreamforce 2019, the luminary spoke on the third day of the event and was no less a personage than former President Barack Obama.
And with Jitterbit’s win in yesterday’s Developer Demo Jam, our Dreamforce is just beginning!
Before we get to our win, let’s begin with the news.
President Obama Talks with Marc Benioff
Dreamforce 2019, day 3 had the featured speaker as Barack Obama, the 44th President of the United States. His speech was widely anticipated, and while he generally stuck to congenial topics, he did sound one note of dissonance. In his fireside chat with Marc Benioff, the former President said:
“So much of the political turmoil we’re seeing right now has to do with people feeling legitimately, materially insecure, because of those disruptions. And we have not adjusted our social institutions to make sure that we benefit from this huge rise in productivity that comes from technology.”
The Future of AI
With the President’s words in mind, here is what was said about the direction AI is headed at Dreamforce 2019. The topic warranted its own keynote called Creating the Future of AI: How Salesforce Research Advances AI for CRM. The discussion was hosted by Chief Scientist Richard Socher.
In past keynotes, Socher had spoken about how AI was driving the Fourth Industrial Revolution thanks to three advances in technology:
- Large Data Sets. Big enough to contain the most important patterns of the data needed to make statistically accurate predictions,
- Better Algorithms. Such as Deep Learning that are trainable end-to-end,
- Faster CPUs and GPUs. Fast enough to process all these massive data sets and iteratively train the algorithms.
Among the many branches of AI, Natural Language Processing seems to have the most current applicability to CRM. At least until now. Most salespeople don’t like devoting time to manual data entry. Logging calls, creating reports, and updating records are not time spent selling. Voice is a more natural and intuitive way to interact with computers. Yet Salesforce Research is not stopping with Natural Language Processing.
At Dreamforce 19, Salesforce committed to extending the use case of CRM from Natural Language Processing to Machine Vision, and specifically to optical character readers (OCRs). There is a lot of corporate data that exists in the form of relational tables. The most memorable part of the keynote was when Xi Victoria Lin, a Research Scientist at Salesforce, demonstrated how a cell phone camera could execute a SQL statement.
The demo was very insightful because the first time Lin tried it, she got an error message. Much to her credit, she cool-headedly adjusted the frame of the picture, and the demonstration was successful. Using her cell phone, she was able to take a picture of a shipping label and enter it into a relational database.
This incident points to a significant limitation of AI in its current state. As long as AI remains narrow, it would be better to think of it as assisting instead of replacing human beings. The AI Lin demonstrated undoubtedly made her more productive. She was able to use her camera as a keyboard without any specialized knowledge of SQL. But the AI only worked because she was there to handle an unexpected event. Without her general intelligence to solve the problem, artificial intelligence would not have operated correctly.
Demo Jam Winner – Jitterbit
Jitterbit demonstrated a bit of problem-solving Wednesday when it won the Developer’s Demo Jam against stiff competition. Demo Jam is a game show-style contest where AppExchange partners show their technology in three minutes in front of a live audience that votes for their favorite demo. A Jitterbit team lead by Matthew Lonergan beat Prodly, Moover, Elements.Cloud, Flosum, and Odaseva. This victory is the kind of problem the current generation of AI would be hard placed to solve. Congratulations Matthew!
Stop by our booth at the #DotOrg lodge at Westin St. Francis hotel or our Integration Concierge booth at the #Trailhead zone to learn how Jitterbit can solve your integration needs.