Home > Computing > From Jeopardy champ, to ace chef: Here’s…

From Jeopardy champ, to ace chef: Here’s what’s next for IBM’s Watson supercomputer

From slaying game show legends to serving up tasty recipes, IBM’s Watson supercomputer has made quite a name for itself. Back in 2011, when it took on (and dominated) Jeopardy masters Ken Jennings, and Brad Rutter, Watson was permanently seared into the public’s consciousness.

Since then, the wonder-system has gone on to make a substantial impact in the world of business, and across multiple sectors as well.

However, in many ways, Watson’s story is just beginning. During a presentation today at its new Watson Group headquarters in the heart of New York City, IBM gave us a glimpse at what’s next for its most famous technological marvel.

Stephen Gold

IBM exec Stephen Gold says Watson is designed to help people “execute their professions better, not just cheaper.”

IBM executives, along with representatives from partner firms who are leveraging Watson’s understanding for natural human language in order to super-charge the sites and services they’re working to develop, gave us a brief look at how Watson could make your life easier sometime in the near future.

Related: Meet Synapse, IBM’s computer chip that can help the blind see, detect diseases, and much more

Though 3,300 start up businesses are developing Watson-powered apps in tandem with IBM here in the Big Apple alone, the showcase highlighted a small handful of firms from disparate sectors of the business world, including customer service, travel planning, and healthcare. These outfits are working on are tapping into the awesome power of Watson to deliver better services to people and businesses alike.

Here’s the problem that these Watson-powered apps are essentially trying to solve. Search engines are supremely useful tools, there’s no doubt about that. However, as soon as you type something into a search engine and hit “Enter,” you’re then tasked with sifting through mountains of data that can range from extremely useful, to downright useless.

IBM Watson 2

IBM says that Watson “can fill gaps in our knowledge.”

Using a Watson app development platform, thousands of firms are building Web apps and services that combine Watson’s ability to understand natural human language with its unmatched ability to sift through endless troves of data, in order to deliver precise, relevant information at the snap of a finger.

The apps these firms are making are in various stages of development, but if there’s one thing we know, IBM Watson-powered apps have the potential to connect people with information in a way that search engines can only dream about.

These are some of the outfits that were highlighted by IBM today.

Wayblazer

Wayblazer is helmed by Terry Jones, who founded such household names in online travel booking like Travelocity, and Kayak. With Wayblazer, Jones hopes to strengthen the relationship between the app and travelers by using Watson’s ability to draw on previously-mined information, along with its understanding of natural language, to deliver answers to prospective travelers in no time.

Here’s how Wayblazer works. When you ask Wayblazer a question, Watson, which works in the background, picks out data from a multitude of sources, including travel sites, social networks, blogs, and others. It combines this with the visitor’s prior Wayblazer history in order to deliver answers to questions that the app thinks that particular user will like.

Wayblazer

The problem with search engines is that they offer “clues, not advice,” Wayblazer head and Travelocity co-founder Terry Jones says.

For instance, if your history indicates that you like barbecue and live music, and you ask Wayblazer to find a place with live music, Watson will draw on all that previously mentioned info to provide you with results that get you the best of both worlds; a place that has live music, and serves barbecue.

By partnering with IBM and Watson, Wayblazer is hoping to put an end to the days of having to use multiple travel sites and services in order to find what you want.

“To date, online trip planning has been a complex and time-consuming chore lacking a way to connect, organize and personalize data,” Jones says. “WayBlazer, makes sense of the information overload and presents it to consumers as a personal travel concierge.  Travel suppliers from destinations and hotels to airlines and rental car sites can use WayBlazer to provide a one-stop solution with personalized recommendations, accelerating the pace and frequency of online bookings.”

Right now, the Austin Convention & Visitors Bureau in Texas is using Wayblazer in order to get better at booking conventions, boost hotel bookings, and more. At the outset, Wayblazer will be a business-to-business service, but the plan is to also allow the average consumer to make it dead simple to plan a vacation at some point.

Red Ant

How many times have you called up a customer service line, or tried to get more information about a product from a sales associate, only to come away from that interaction with your questions still unanswered? Red Ant, a firm based in the U.K., is developing a Watson-powered app with hopes that it makes these kinds of experiences less commonplace.

Red Ant’s Sell Smart app is designed to take note of every individual customer’s purchasing preferences. When a user asks Sell Smart a question, Watson works in the background, tapping into information related to buying history, purchasing wish lists, buyer demographics, product info, customer reviews, and more, in order to deliver an optimal shopping and service experience.

Red Ant

Red Ant says that the idea behind its Sell Smart app is to provide access to “knowledge that can take sales assistants months to acquire” which can be had “in seconds.”

Sell Smart features a simple interface that you can use to ask it questions by typing them in, or using your voice. Red Ant hopes that, by developing an app that’s powered by Watson, fruitless encounters with sales and customer service reps will one day be a thing of the past.

LifeLearn

One of the hardest parts of figuring out what’s wrong with your pets is the fact that the poor creatures can’t verbalize what’s wrong with them. If whatever is afflicting them isn’t externally visible and obvious, it could take multiple trips to the vet, along with loads of time and money, before whoever is working on your buddy to figure out what exactly the deal is.

LifeLearn hopes to solve that problem with Sofie, an app it’s developing that’s designed to help veterinarians diagnose what’s wrong with a patient faster, and more efficiently. Powered by Watson, Sofie mines data from textbooks, teaching hospitals, and veterinary professionals in order to answer a doctor’s questions using a simple text-based interface.

Sofie vet app

LifeLearn hopes that its Sofie app will help vets solve cases in less time, thereby improving the quality of patient care, while cutting costs as well.

With Sofie, LifeLearn hopes that a vet’s experience with it can grow to be the equivalent of tapping into a “third or fourth” professional opinion. Right now, the Sofie app is restricted to figuring out what’s wrong with dogs and cats, but LifeLearn plans to add other types of animals to the roster as well, including horses, birds, and other critters.

Challenges posed by developing apps in tandem with Watson

While all of the ideas above sound good on paper, the strength of those services will only go as far as the developers of these apps want it to.

Contrary to what you might have seen in those Jeopardy matches in 2011, Watson doesn’t automatically know everything about everything, and this goes for its abilities to excel in these kinds of applications as well. In the case of Red Ant, LifeLearn, and Wayblazer, when developing these apps, Watson must essentially be taught to answer the questions of anyone who uses them, while also providing razor-sharp information.

For example, as we stated above, when a vet asks Sofie a question, Watson needs to sift through mountains of information to deliver an appropriate response to the question. When we spoke to representatives of LifeLearn, we were told that all of the information that gets fed into the app (and by extension, Watson, needs to be vetted (no pun intended).

Watson Food

During a break in the action, IBM served up some food whose recipes were concocted by Watson. On the left is a helping of Scandinavian Salmon Quiche. On the right is a Dutch Brazilian Vanilla Croissant with a Porcini Mushroom glaze. When it comes to the kitchen, Watson has not lost its touch, or its flair.

In Sofie’s case, all of the text and information that Watson gets to draw on gets uploaded by the app’s handlers. Therefore, it’s the strength of the information that Watson has access to which will play a big role in determining how effective these Watson-powered apps will be when answering a user’s question.

On top of that, whenever you ask a question using any of these apps,  they’ll often give you a relevancy, probability, or accuracy rating. If you use natural language to pose a query, and the rating is low, the app’s handlers will be able to take a note of that by looking at the app’s performance history data. In cases where the apps spit 0ut low ratings, it will be up to the developers to help Watson get better at solving problems and answering questions that it has trouble with.

In other words, this is all part of a teaching process. After all, Watson is a learning computer, in the truest form of the phrase.

There’s much more to come from Watson

As VP of IBM’s Watson Group Stephen Gold puts it, Watson is designed to help people “execute their professions better.” Will these Watson-powered apps deliver on that promise? Only time will tell.

In the meantime, the apps and services which IBM highlighted today are just a small sampling of what Watson will be working on next.

It will be interesting to see if Watson can dominate industries like healthcare, travel planning, and customer service the way it mopped the floor with a couple of Jeopardy masters.