When IBM’s Watson technology isn’t winning Jeopardy challenges, it’s been talked up as a potential medical assistant, helping doctors diagnose health problems. But IBM also sees Watson as a potential smart personal shopping assistant, helping retailers answer a wide array of consumer questions about products and recommendations.
IBM is at the National Retail Federation Annual Conference and Expo in New York showing off how Watson can help retailers get a lot more intelligent and responsive in dealing with consumers. IBM believes that Watson could serve as a phone customer service line, in-store kiosk or as a resource available through tablet-wielding store clerks. The idea is that consumers will be able to ask all manner of questions about products, trends and recommendations and Watson can come up with an instant answer that pulls from millions of data points, both structured and unstructured data.
A consumer could ask for a gift recommendation for his 7-year-old nephew and Watson would be able to canvas sales data, supply chain records, product information, social mentions and other content to come up with insightful answers that could be extremely relevant for that user. Or a shopper might have questions about what’s popular now or what particular product model is right for him. Watson would gather all the data and weight it all to give it a level of confidence in the answers it provides.
Watson’s strength lies in being able to use natural language processing to understand a consumer’s real intent and put those cues into context to extract the right information from various data sources. And then it gets smarter through machine learning. With so much data created each year and products evolving constantly, even store associates are hard-pressed to answer all the different queries of consumers. Watson could come in and help direct consumers to firm answers that could ultimately lead to more sales. Consumers are already consulting multiple sources and are using mobile phones to conduct in-store product research. Providing Watson as an option could help keep consumers engaged and spending in-store or through a retailer’s online channel instead of fleeing to a competitor.
Retailers would still need to tailor Watson to their businesses. A company might want to make sure it can direct consumers to product it actually stocks, instead of pointing them to competitors. But if done well, it could provide a new level of interactivity with consumers. And it could provide some opportunities to up-sell consumers on other products.
“Building Watson was about building the technology that allows us to better understand that natural language dialogue problem,” said David Ferrucci, IBM Fellow and principal investigator of Watson, during a panel discussion Sunday at NRF. “It’s not an easy problem. It’s not about looking things up.”
IBM representatives I talked to at NRF have no time table for when merchants and retailers could start using Watson. And it’s likely that IBM will have to improve and refine the algorithms it uses to be effective in consumer retail. But I think we’ll continue to see much more of these smart artificial intelligence technologies applied into a wider and wider array of settings. It’s still early but there are a lot of verticals that can use a Watson approach. IBM thinks that financial services, government, manufacturing, fraud detection, are some of the sectors that could tap Watson’s intelligence.
As we’re seeing glimpses of with Siri, intelligent systems that can understand natural language can create really engaging experiences. Systems like Watson and Siri still need to deliver the goods but there’s no reason why they can’t be applied in all manner of environments.