Or listen in your favorite podcast app

Apple Podcasts  /  Google Podcasts /  Stitcher

All Things Einstein and Voice Technology

The crew who work on Einstein voice technology for Salesforce dropped by Mission HQ to discuss unstructured data, give a crash course in deep learning technology and what tech they are excited for in the future.

“People have a misconception that voice technology is simple, but there’s a lot under the hood and there’s so much you can do with it.” — Ceverly Strand of @salesforce ‏#ITVisionaries

Show Notes

Michael Machado (LinkedIn), Senior Director of Product for Einstein Deep Learning Solutions, Ceverly Strand (LinkedIn), Product Marketer for Einstein, and Julien Sauvage (LinkedIn), Head of Tech and Product Marketing for Einstein, all stopped by Mission HQ for this special episode of IT Visionaries!

This Salesforce crew is part of the team that is bringing Einstein and voice technology to customers everywhere. They took some time away from the lab to discuss with Ian all the ways Salesforce is doing innovative work in this exciting field. They also talk about what is going on in the AI marketplace, explain what deep learning technology actually is, and take a look into the future of tech.

Topics Discussed: Voice technology, deep learning, AI, IT, unstructured data.

Introducing the roundtable — (1:30)

• Michael came to Salesforce three years ago from a company called MetaMind, which was focused on deep learning technology. Now he is bringing unstructured data into Salesforce and launching voice technology.
• Ceverly works on marketing the unstructured data, as well as vision and voice software.
• Julien does marketing for all of the Einstein products Salesforce offers.

What’s going on in the AI marketplace? — (3:05)

• Michael: More and more companies are interested in bringing unstructured data into the enterprise.
• Unstructured data — video, texts, messages over the phone and voice itself (aka streaming audio) — used to be very difficult to analyze. But deep learning and AI have provided a way to analyze the data and make it structured in a new way.
• “In certain ways, there are vast amounts of data being collected for no reason at all, but some people are looking at that as a potential strategic advantage for the future. I think in the future we’ll be leveraging that data in a more trusted way and in a way to deliver new experiences for users.” –Michael

Defining “unstructured data” and using it to make a change — (5:10)

• Images, blocks of copy on surveys, and more inputs that are not specifically defined for analysis are considered “unstructured data.”
• Using tools like keyword extraction help aggregate unstructured data, and the tools are becoming more advanced to work at a much larger scale.
• Julien: There are four main benefit statements when it comes to AI and use cases around AI — discover, recommend, predict and automate.  
  • Discover means to discover hidden patterns in data.
  • Predict is at the core of deep learning and AI and helps to focus on leads or accounts that might convert to sales.
  • Recommend is used to recommend the next best action or engagement to use.
  • Automate allows for the automation of workflows with AI.

How Einstein works within the four benefit statements — (7:55)

• Einstein helps sales reps know what opportunities to go after and recommends which leads to pursue. It can also be used to create a customized experience and predict when to reach out or make the next move in business or sales.  
• “The jury is still out on AI. We believe people don’t buy AI just for the sake of having AI. People buy AI for their CRM because they want their CRM to be smart, to be automated, to be intelligent and to be AI-powered.” –Julien
• “Einstein is everywhere in our narrative and in our customer success stories.” –Julien

A crash course on deep learning — (13:30)

• Many of Salesforce’s products are built directly into the platform already.
• Think about your business process and what data you have. What do you want to achieve? How will you use the data to achieve what you want?
• You are not looking to make employees or processes one percent more efficient. You are looking to create an environment or a system that will yield the absolute best possible results every time. Sometimes that will mean taking tasks off of their plate that can be automated or analyzing how they succeed in some areas and fail elsewhere, which is what deep learning helps with.
• Identifying small, incremental solutions by using deep learning and then implementing those solutions can have monumental effects on productivity.

The excitement surrounding voice technology — (20:45)

• Ceverly: This is about getting better data into the system.
  • • Agents in the field will take notes by hand and then never get that data into the system for others to use or get insights from. Or, if the data is input, it is not always complete.
• “Voice is helping change things by allowing you to put great data into the system at the moment.”
• Michael: Different companies have different thought processes about how to leverage voice data.
• Privacy and trust will play a huge part in finding the best use for voice technology.
• “We’re going to start seeing voice as a primary input in devices you’ve never even thought of.” –Michael
• Analyzing past data is great, but there’s more to it than that with Einstein — voice is going to be the new UI.

What everyone is most excited about for the future (31:45)

• Ceverly: Expanding the voice use cases.
• “People have a misconception that it’s a simple technology but there’s a lot under the hood and there’s so much you can do with it.” –Ceverly
• Michael: Autonomous Assistants and the relationship individuals build with them.
• Julien: More data and the possibilities of improving technology with it. It’s a virtuous circle of data collection and improvement.

Customer success stories with Einstein Voice — (36:20)

• A workers comp organization has been using Einstein Voice to connect customers with providers, improve customer experience, streamline the process and get workers back to work faster.
• Ian: “If you’re not creating frictionless experiences, your customers are not going to get the services they need.”

Lightning Round — (39:00)

• Favorite apps: Twitter and Flipboard are helping the panel stay connected.
• Hidden Brain, sports podcasts, Joe Rogan and French fiction novels and Riad Sattouf books are the top content the panel is consuming.
• Favorite vacation spots: Charleston, Northern California and anywhere in the Mediterranean.