IBM’s Watson cognitive computing platform is almost a victim of its own fame. Watson famously beat all-comers to win the Jeopardy game show a few years ago. At the time, the general public (helped, it has to be said, by IBM’s marketers) assumed the win was an indication that, in short order, smart computers would be everywhere and intuitively making the right decisions in every situation.
+ Also on Network World: IBM’s Watson wants to do your tax returns +
Since the Jeopardy win, however, IBM seems to have had a hard time finding good market fits for Watson. This isn’t a criticism of IBM in any way. The reality is that while many consumer brands apply artificial intelligence (AI) to their products (think Amazon book recommendations, Google maps smart routing or Apple’s Siri) most existing examples have been from those companies building AI tools themselves. There are far fewer examples of enterprises leveraging a third-party cognitive platform to build into their own applications.
I sense a change is coming, however. Many other technology vendors have started commercializing their own internal machine learning tools. (Yes, I realize that there are differences between cognitive computing, AI and machine learning. But at a very coarse scale, and from the perspective of this article, I will use the terms interchangeably.) Google, Microsoft and Amazon Web Services have all opened up parts of their own machine learning platforms for other organizations to use—whether it be sentiment analysis, image recognition or pattern matching.
Invoca’s Watson-powered Voice Marketing Cloud
So in light of the slow, but steady adoption of these tools, it is interesting to hear from Invoca about its use of Watson. But first a small intro to Invoca. The company offers a platform that allows marketers to gain insights into the customers calling them. Invoca, in something a little bit like Big Brother, analyzes customer calls to indicate why they are calling, analyze what is being said in conversations and generally understand the voice aspects of conversations. The company has an ecosystem of integration partners, allowing it to “inject” call awareness into broader customer journeys.
Invoca teamed up with IBM and started using Watson to deliver extra value from all those phone calls. It uses the Watson Speech to Text API to transcribe phone conversations into text, which are then analyzed to reveal any insight that might be relevant to a marketer or their business. During the conversation, audio streams through Watson in real time, transcribing the conversation and documenting the time any word or phrase is spoken. Invoca’s algorithm then takes over and identifies keywords and phrases that marketers classify as “signals” that indicate specific actions, such as intent to purchase.
How Frontier Communications uses Watson-powered Invoca
This is an interesting example of Watson being applied to a technology problem, but we can take that one step further and see how the combined solution is being utilized out in the wild. Frontier Communications uses Watson-powered Invoca to identify spoken phrases in real time that are taking place between the caller and the agent to determine the final outcome of the call, along with other caller intents.
For example, Frontier is interested in understanding the different types of conversions that happen over the phone, so it is defining specific intents in Invoca’s system, such as whether the agent mentions “confirmation number” or “installation date.” Frontier then uses the data to understand conversion rates across multiple channels and optimizes its marketing spend.
“In the telecommunications industry, as for many other considered purchases, it’s common for people to research online, but then head offline to make a phone call as they get closer to purchasing,” said Chadd Ryan, senior manager of Insights at Frontier Communications. “By working with Invoca, we’ve been able to analyze and act on a vast amount of data that traditionally lived only in our call center. And with the power of Watson’s machine learning, we’re able to get an even deeper level of insight from spoken conversations that we can immediately optimize against to improve the customer experience and drive revenue.”
Overall, it’s a nice, real-world use case for Watson and, more generally, proof that these tools are being used out in the wild.
This article is published as part of the IDG Contributor Network. Want to Join?