Dialpad Execs Discuss R&D, AI, Changing Partner Landscape

Last month, Dialpad announced a $50 million investment into the continued research and development of artificial intelligence-related technologies. The goal: to accelerate the development of features to automate business processes and provide predictive insights.

In the next five years the company wants to take AI “to the next level” said Craig Walker, founder and CEO.

For Dialpad, that may not be an audacious task.

The company disrupted the collaboration market years ago, company officials said, by developing and releasing the first real-time speech recognition engine for enterprise conversations. Dialpad has since extended its AI capabilities by delivering industry-first features such as real-time assist (i.e., suggested answers to questions), virtual agents (AI-powered self-service web and chatbots), and AI CSAT (i.e., inferred customer satisfaction from service conversations), among others.

Dialpad’s Mike Kane

In this interview with Channel Futures, Dialpad’s Mike Kane (global channel chief) and Dan O’Connell (chief strategy officer) discuss the process of developing AI for their brand. They also talk about how the fastest growing part of the business is the contact center. Finally, they delve into how the role of partners is changing and what that means for the company’s relationships.

Dialpad’s Dan O’Connell

Channel Futures: How do you differentiate yourself from your competitors?

Dan O’Connell: There are really three technologies I would classify when people kind of talk about AI. The first one would be automated speech recognition, which would include how one gets transcripts. It all starts from there. We do our own speech recognition in-house. We do monthly benchmarking against pretty much other speech recognition providers out there. We outperform all of them in terms of accuracy. People that use Dialpad know that they’re going to get the highest level of accuracy in terms of just what a transcript looks like.

The next part gets into natural language processing. When it comes to sentiment analysis, we infer customer satisfaction from conversations. All of those are proprietary models, meaning those are built specifically on Dialpad data for specific users of Dialpad. We’re doing models that are designed for health care businesses for measuring customer satisfaction. We do different models for businesses that are in finance or ecommerce. One of the biggest differentiators for us is we do all of this in-house; we’re not using third parties for it. They’re not generic models. They are built and trained on specific use cases.

And then the third technology gets into what’s called semantic search, which is how you essentially retrieve information from a knowledge base and present it for one use case. The key differentiator for us is, we don’t partner with people. We think this is fundamentally different. We think it’s really, really important to do this in-house. And we think it’s important to do it in-house because you can control your own destiny and you can derive better accuracy. Better accuracy means more value for customers.

See our slideshow above for the rest of the conversation with Dialpad.