AI: The Future Is Here

Damion Heredia, IBM Watson

Damion Heredia
VP of Watson Strategic Partnerships, IBM Watson
View Slides
Damion Heredia is the Vice President of the "With Watson" business across IBM. He is responsible for strategic partnerships with ISVs (e.g. Slack, Twitter, Salesforce, etc) using Watson technology in their products and services to differentiate in the market.

Damion Heredia was previously head of Products & Design for IBM Bluemix, the world's largest Cloud Foundry based platform-as- a-service. Damion also launched IBM's MobileFirst software strategy and led the acquisition of Fiberlink MaaS360, the leading Mobile Device Management platform. Previous to IBM, Damion was VP of Product for Lombardi Software and Trilogy Software in Austin Texas. Damion holds a BS in Electrical Engineering from Purdue University.

Video Transcript:

Great venue, looks like a great crowd. Thanks for having me. You get to endure me for the next 20, 25 minutes. First off, my name is Damion Heredia. I'm a vice president within the Watson product group. I focus on what we call the With Watson program in business. It's helping partners and clients embed our technologies and their products and services to take to market. You may have seen that logo you see on the screen there and the Superbowl ad with H&R Block for example, who uses Watson to teach the tax code and help advisors and other partners like that. We're really proud to be part of the Spigit community today and to be here to talk to you about AI. As we dive into it, I'm going to be talking about AI. I'm going to define it a little differently than most people today. We're going to go through why it's so important, how it's being used, where it's going and how we tie this all back to the innovation.

But first what I want to do is kick off with a commercial, the IBM ad scene that we launched a little bit ago, a couple years ago actually, around Carrie Fischer hosting a group therapy session with a bunch of robots. Play the video.

First, from a media standpoint, when somebody hears the letters, AI, put together, they think evil machines, things are going to take over the world, domination. That's pretty much anyone you ask. You're going to think about artificial intelligence as the source of that technology. But the way IBM views, and I think the way our business are starting to view it, is it's really about augmented intelligence. It's not about replacing humans, it's about helping them make better decisions. It's helping them learn, reason, understand more information to do their job better. To scale their efforts, to go through their limitations, and to really provide a different level of intelligence that they can use to do better in their business in their daily lives. That's an important differentiation for us. We don't think the goal is artificial intelligence, it's augmented intelligence.

Let me talk a little bit about the history of that. AI, and augmented intelligence is not new. Humans have been fascinated with the notion of using technology to scale their efforts, their skills, their strengths. Rather than be limited by lifting a few hundred pounds, we invented in 1839 the steam engine to lift 16,000 pounds. That's augmenting our skill as humans, right? It goes further. Humans look out into the horizon, and were not satisfied with just seeing a few mere miles into the horizon. We want to see the universe. We invent telescopes that see out 15 billion light years into space, actually into the past. For us, in fact I can make statement that you can't find a useful technology that isn't built on the premise of helping us scale our human attributes and strengths. That's what it's about. It's what you're innovating around.

For us, at IBM, in 2011 we started a journey through, the cognitive journey. We felt the center of that problem was data. Massive amounts of data are now available to us that weren't available before. It's an explosion of data that we've never seen before in our industries. To make sense of that data is very difficult. In fact, about 95 to 98% of the data you have in your enterprises is unusable in most machines. It's called dark data. It's stuck in videos, and images, and handwritten notes, and PDFs, and contracts, and wikis. It may be structured in terms of language, but it's completely useless to most machines. You can't query it and get an answer back of what you want. We as IBM started investing in a cognitive platform we called Watson, a machine that could ingest large massive amounts of information to understand that information in a way we could use it. We also then developed a way to reason. To make decisions with you and to do recommendations on incomplete data.

But we have to then make a leap of assumption and that reason to something that a normal machine can't do. Then we had to teach it how to learn. How to teach it, you know, every article ever written about a particular topic, every medical journal, all literature. Anything that you could get it hands on to learn and to grow its, what we call, machine IQ or people will refer to it as MIQ. The more it learns, the better it gets, and on any topic. Then we had to teach it how to interact. This goes beyond just robots and speech, but it also is the way we interact through chats, through texts, through motion, through tone. We did research on robots and found out that if you have a robot sit in the room that if it's just standing there, it actually wigs people out a little bit. But if it moves its arms a little bit and twists its head and acts like you know, it motions that it's listening, it feels more human like. It's more engaging. You're more likely to trust its recommendations if you feel it's more like you. It's not replacing you, but it's helping you make decisions. That goes through a bunch of examples we'll go through next in your businesses.

For us, the data is the center of any journey through cognitive. We had to build a machine for that. Most people know that machine as the one that won jeopardy. When you say, "Watson," they think jeopardy. We built that for one use case, and A and Q, right, not Q and A. You give us an answer and we give you a question. But it could do that in record time. It could interact, it could reason, and it understood and learned constantly. That was the genesis of the cognitive platform for IBM. But we wanted to take that further. We really wanted to drive augmented intelligence for all our customers and businesses and for the world. We broke down the Watson components, deconstructed it into smaller API cloud services that are available now to the public. Now anyone can come in, swipe a credit card and use the APIs that we built our augmented intelligence capabilities with for our clients. You can build them. Now these were geared toward the developer, because a lot of your enterprises they'll tell you the developers call themselves the king makers, right? They own the compiler, they own the systems, they own the apps that you want built on the phone. By enabling them to have better tools to be able to bring that augmented intelligence and to be able to make sense of all that data, was an important strategy for us.

Now it lead to a wide range of use cases that by ourselves we could imagine, but by putting those type of tools in the hands of developers and business people and innovators from both sides, you're able to solve things like help doctors prepare the appropriate treatment plan for cancer patients. If you're a 26 year old female in Brownsville, Texas that has lung cancer, you're pregnant and you've never smoked. What clinical trial and treatment plan is best for you? There's no way that doctor can know all 6,000 active current clinical trials in the world going on right now. But Watson does. It can recommend, "Here are the top three that you should look into as a physician of caring for this patient, and also based on the genome, here are the type of medications and treatments that we would recommend that you need to go further into the next steps and test out." The doctor makes the decision, not the machine. But it's helping the doctor understand what's available to him and to be able to use that to augment their experience.

We train Watson on oncology, pathology scans and results, x-rays, anything we can get our hands on to train it to make that MIQ go up is advantageous to then provide it a better augmented intelligence. Okay? That's just one industry. If you go forward, what's happened over the last couple of years is a boom of using AI and augmented intelligence in everyday businesses. It's not reserved anymore for the elite. Used to be thought of AI as, 'Oh, that's for those companies up here that have that pot of gold they can spend on these really high end problems.' It's being used in every industry. Very soon, you won't be able to really drive your innovations without these type of notions built in. Whether it is understanding the law better, to debate one side of the issue or another, the way you drive, the way you educate your children, to hospitality when you walk into a hotel room or a cruise ship and having a digital concierge interact with you and know that's date night and you should go to this restaurant versus that one, to retail and shipping.

For example, we use weather as part of Watson to drive a lot of insights. We found out through a retailer that sells games, video games, that if a snowstorm is approaching a certain region of the US, they can ship out the latest titles, redirect shipments to flood those markets with more games than they would normally do that week because school usually gets canceled and guess what kids do when they stay home from school? They run up the store, buy the video game, come home before the snow storm locks them in and they play the game. That's just one of 100 things that retailer found, relationships between all this massive amounts of data that then can help you make better decisions in retail. It goes on through all industries.

Let me talk to you about a few of those and how we see those trends evolving around innovation. One of those is in banking, in RBS. Customer care is a big area for AI, because it's a way to reduce cost, but also really it's about scaling the experience to make a better experience for your users, and getting a better turn on customer sat when you're engaging. If you want to set up a new account, you may say like, "What type of account would you want to set up?" They may respond back with savings or checking. That would be a pretty easy chat and experience to go through. You know, it's almost pretty simple. Ask a question, get a response. But it's much tougher that because the user may respond back, "I would like an account that has a high yield, but low risk, and I can cash out anytime." The way they describe that to you most machines wouldn't be able to answer. They can over 4, 5, 6, 12 questions deep. We understand the entire context of that conversation. It's not just a question and answer anymore, it's a conversation around a set of products. We've ingested all literature about every banking product, all the support tickets, every piece of information around those particular product services in public demand, private demand in your company to be able to answer that question better.

Every time we do it, we learn more and more and it auto improves itself to make a better experience for your customers. It also can seamlessly hand off to an agent on the other side to go work that issue. That's really the augmented piece. It's helping bring in customer care solutions. Yes, taking care of tickets automated way, but then handing off to agents to be able to handle those better for you and it has more context. As the agent's working it, they have their own AI next to them of Watson telling them how best to handle this particular question. It's really about scaling your employees, right, and what they can do.

Now one of the new concepts also that's coming about, is a notion of a cognitive profile. Understanding your end user's personality, location, what they're doing, what they buy, how they use, where they drive, anything you can get to what they on the internet and Twitter, to the writings they've done, anything to be able to build a profile on that particular user so that you can help them have a better experience of your product and service. Here's GM, and OnStar start using it, to drive an experience in the car itself of interacting with a particular errand the person is doing, and being able to understand the behavior and traits and personality of that driver in that car at that time. For us, it's all about the data, but then it ends up in an experience the person can interact with.

Next is a company that we work with that built their company around Watson. They're called Influential Co. They're based in LA. They said, you know, "What if I could apply Watson to all the social data out there, the big fire hose on Twitter, etc. and basically find all the top influencers based on personality?" Watson understands the five big attributes of personality. There's 27 different attributes total inside there. It understands if a user is curious, excitable, if it's hedonistic, and what level and we can rank them based on what they're doing in the social media. If you're in this case, Kia or if you're the example last month, two months ago in South by Southwest was Mazda with the CX9 release, their new campaign around the CX9. They want to make sure that their message around CX9 gets applied to the right influencers in social media to scale their efforts. They want to match an excitable, curious message with people who are curious and excitable. They don't want those to match. Only Watson's able to understand the personality of that side and be able to make that targeted message happen better.

Now I'm going to use an example internally. How many of you have the Weather Channel app on your phone? We get around 4.4 billion forecast requests a day. The Weather Channel's part of IBM. You say, "Well why is it part of IBM?" At the center of the Weather Channel, is data. It's data from 2.2 million sensors around the world updating every second and pulling it into our cloud to process, so that we can understand insights from that data and then deliver that back out to customers to make decisions for rerouting planes, to shipping logistics, to hail storms coming in through an area telling the insurance providers to text you to move your car inside. Anything. We also build a chat experience where you can go into the app and have a conversation with the Weather Channel app and talk about what you're looking for in a forecast, what do you want?

But it wasn't exactly where we wanted to be. It's a great kind of notion to interact. What was interesting about it, is when we thought about how to change the entire business model around advertising. A normal ad in the Weather Channel app is a static image of, here's an ad for Theraflu. What we did, is we made it different. We said, "What if we put up an ad around just the generic clearing up cold and flu questions, and you can interact with that ad in real time. You can ask it, what's the difference between cold and flu? What are the best remedies? Which herbal medications should I use? How long does it last?" It will understand the intent of what you're trying to do and respond back. We've leveled up from an ad, a static image of impression, to a totally different engagement model now with Watson Ads that you can interact with users in a different way. That's changing business model. We talk about innovations, there's a range of the different types, but definitely when you can rethink business models, that's a significant impact on market.

For you today and just a fabulous opportunity to have thinkers combined in one area here around innovating, the question for you is, how are you going to innovate in a world with AI? How are you going to bring that augmented intelligence to your employees, to your customers, to your suppliers, whatever it may be, in a way that differentiates you and drives more for your business? Thanks for your time. Appreciate it.