Omilia’s head of North American business development Quinn Agen discusses the use of Omilia NLU Speech Technology in the enterprise call center as part of the Enterprise Connect 2018 Innovation Showcase.

My name is Quinn Agen. I’m responsible for business development at Omilia. We offer a virtual assistant platform for conversational customer care on all channels, from the IVR, all the way down through smart speakers, web-chat, SMS, email, and social media. We have deployed our dialogue management platform since 2010, starting with Vodafone in Greece. We currently have 24 production deployments of our dialogue management platform in the enterprise call center, as well as web-chat channels. In North America, we’ve been live for over 18 months at one of the biggest financial organizations in Canada in their call center, servicing all customers from both Canada and the U.S.

Principally, we work with financial institutions, telecom operators, and insurance companies. I would say the interesting thing about the platform for the enterprise is that it offers the business the opportunity to design their virtual assistant application once, and deploy the same application across all channels, leveraging not only the design and the voice-user interface, but also all of the back-end integrations, CRM, data links, as well as your business logic for authentication.

Why is speech important for enterprise communications and collaboration?

So, our platform, basically what it provides is a conversational, unstructured user interface for customer care. There’s no predetermined flow in terms of, you know, “Say this to do that, say ‘cancel’ go back to main menu.” It’s the customer that drives the conversation in natural language. In the voice channel, it’s extremely important because it’s much more demanding, voice, than a text-based channel when you’re talking about web-chat, you know, a session can…someone starts a web-chat and they go away to get a drink. When you’re dealing with a voice-based channel and it’s over the phone, even a couple milliseconds of delay in response can lead to a poor customer experience. And so therein lies the importance of having a real-time speech-to-text transcription as well as deep NLU for really understanding the true intent.

What really makes an unstructured conversational user interface possible is the ability for the platform to maintain context and memory. You know, if we’re having a human-like conversation, and you refer to something that we were talking about five minutes ago and I can’t remember what that was or what we were referring to, we’re not going to get very far in our conversation. The same rule applies for machine-driven conversations and that’s really the holy grail to providing an unstructured conversation user interface. And we do this with Omilia’s dialogue management platform, DiaManT.