Interfaces built on voice control and chatbots may still feel overly futuristic for some consumers. But even if you resist the notion of having conversations with your computer, natural language processing technology is still changing the way you interact with the world around you.
If you’re looking to buy a specific product, for instance, you’re probably more likely to search for it online than to dig through the product catalogs of multiple retailers.
The reason why searching works better than browsing is that computers are becoming increasingly skilled at deciphering our plain English queries. But as the segment of consumers who expect seamless natural language interactions grows, so too does the pressure for retailers to get with the program.
Addstructure, a Chicago-based natural language startup, works with retailers to take the fight to e-commerce titans like Amazon. The company’s suite of natural language processing technologies helps retailers capture organic search traffic, increase the precision of on-site searches, analyze customer interactions and enable voice search capabilities.
Unlike Alexa, which is specifically designed for ordering Amazon products using Amazon devices, Addstructure’s voice search features are not tied to any particular device or operating system and can be sold to any online retailer who wants to provide a similar experience.
“We plug into pretty much every conversational channel, so you can imagine doing this through OK Google, Siri, an Alexa skill or a Facebook Messenger bot,” said co-founder Will Underwood.
Although voice interfaces have come to dominate much of the conversation about natural language processing, Underwood said Addstructure’s search engine optimization technology is also attracting a lot of interest from retailers.
“Probably the primary place people are using natural language right now is organic search on Google — if they’re researching the best headphones for running or good cell phones for senior citizens, they’re turning to Google,” said Underwood. “We see it as a continuum. We try to be end-to-end natural language, and wherever that can occur, we want to be there.”
Underwood got his start doing natural processing as a graduate student at the University of Illinois at Chicago, where he was developing technology to search for restaurants by subjective attributes mentioned in customer reviews, like “best pad thai,” “best service” and “best music selection.”
Recognizing the potential for other applications, Underwood launched Addstructure with co-founder and fellow University of Chicago alumnus Jarrod Wolf in 2014, and soon developed a partnership with a major consumer electronics retailer to power its e-commerce search engine. Leveraging the same principles as in Underwood’s restaurant project, the engine let consumers search for “quiet dishwashers” or “TV’s with good sound quality.”
The company participated in the Minneapolis-based Techstars Retail accelerator this fall. Upon graduating, Addstructure announced that it has entered into a partnership with Target, which sponsored the accelerator program, to run pilots of all three of its core natural language products. Wolf said the experience of working with Target has been a huge boon for Addstructure.
“The Target teams were really open to sharing information, and their search team shared a very, very large data set with us,” said Wolf. “They’re really open to innovation and working with startups.”
Currently based out of New York, Wolf will be moving back to Chicago this winter to join the rest of the Addstructure team. The company is also in the midst of raising its next round of funding, and will be growing its team from six to at least 10 over the next months.
Image via Addstructure.