Talk is No Longer Cheap - Indix



Talk is No Longer Cheap

“Alexa, play some Christmas music.” Music plays. “Alexa, stop. Play “Let it Go” on Spotify and turn on the living room light.”

These are the typical commands my daughters are directing towards an ever-obedient and patient Alexa, Amazon’s artificial intelligence (AI) powered assistant on the Echo.

Whether it’s Alexa, Siri, Google, Cortana or the newest darling on the AI-powered personal assistant block, Viv, given the amount of attention and media coverage this category of technology is getting lately, we’ve all either heard of them or directly talked to them ourselves. Heck, there are even movies about strange loners falling in love with these intelligent bots. You know it’s mainstream when Hollywood has latched on.

Most pundits watching this kind of stuff believe that these bots represent the next major leap in computing and subsequently human-computer interactions. In other words, they believe this will fundamentally and forever change how we humans do many of the things we do today.

shutterstock_114710245While the full scope of that shift is too large for anyone to write about, today I wanted to share a few thoughts on the impact of such bots on one of the most pervasive things we humans do today – finding information aka search.

Whether it’s searching for directions, restaurants, people, interests or products, billions of people are collectively conducting trillions of information queries throughout the world across devices. In fact, check out the full list of voice commands that one could issue to Google here.

With continuous advances in AI, these bots today already yield better and more relevant results than the typical query typed in a search box.

But don’t just read this, I’d recommend you try this on your own. Try typing in a query in your favorite search engine to find the closest four-star Italian restaurant that serves risotto and is open right now.

Fuhgeddaboudit. You’ll have to wade through text-based search results to figure out the answer, if at all possible, and it will take a long time.

Now try it using your favorite personal assistant app.

The same query, when spoken into Hound (another great speech-based search app), gave me a list of relevant restaurants with Barolo at the top of the list. This one happens to be my favorite Italian restaurant in Seattle and does a great risotto.

While searches for places and points of interest are handled increasingly well by these bots, product search, an area of obvious interest for Indix, still remains almost completely unaddressed by all of these providers.

A product-oriented query like the one above might sound like this: “find me a four-star rated infant stroller that is priced between two and three hundred dollars and on sale at BabiesRUs.”

In order to process this query and produce relevant results that can be read out by the bot, the following high-level steps must occur:

  • Automated Speech Recognition (ASR) must first accurately translate the spoken command to text – most bots do this very well today
  • Natural language processing capabilities must be applied to infer the stroller price range, category (infant) and store – also done fairly well today
  • Named entity recognition must be applied to infer that the store is BabiesRUs – a lot of room for improvement here still
  • A structured query must be put together using all the inferred entities and information
  • The query must be conducted against a large aggregated and structured database of product supply that can yield relevant results
  • The response must be structured in a way that can be parsed and presented back in both visual and spoken methods
  • Users would then be able to drill into each result and see more or transact on the product

Surely this is not a trivial set of tasks. However, the companies involved in this space seem to strongly believe in the value of enabling Conversational Commerce. They are at various stages of evolution on the capability spectrum due to the complexities involved.

At Indix, we are of course very excited about this space and believe we can be a major contributor. We’ve been on a mission to aggregate, structure, and enable access to structured product information for such innovative use cases and it’s great to see that the talk about speech-based commerce is finally materializing into real experiences.

  Download the Pervasive Commerce White Paper
Get Demo

Leave a Reply

Your email address will not be published. Required fields are marked *