I joined my call with Avaamo founder and CEO Ram Menon with one conversation in mind and left with a very different one. I am interested in how Avaamo introduces artificial intelligence (AI) to non-technical users. They have a tool called Flow Designer for building conversational interfaces powered by AI, and I wanted to know more about how non-technical people were learning to develop "bots" via this platform. Ram quickly shifted my perspective, and our conversation, to a higher plane. We instead talked about how to design technology to work for humans, rather than the other way around.

Privileging the Humans in the Interaction

Ram's response to my question of how people are learning to build AI tools (yes, non-technical people are doing that and I'll return to the topic in a future post) was that he thinks of it differently — which means his company approaches it differently. Rather than a conversation about teaching people to work with new technology, we had a conversation about how Avaamo works to let people do things they already know how to do: have conversations.

Avaamo is building fundamental AI technology across a broad area of neural networks, speech synthesis and deep learning to make conversational computing for the enterprise a reality.

I tend to ignore chatbots that pop up on a screen. These are the "How can I help you" buttons that appear on many websites today. My general chatbot interaction feels cumbersome. My assumption is that the jerky nature of the interaction comes from a human customer service representative managing a variety of different conversations at once. I respond by multitasking while still trying to be polite, engaging in hello-how are yous, but trying not to waste time watching a blinking "Heather is typing" notation. 

Avaamo is on the leading edge of improving these interactions by creating systems that take on much (all in many cases) of the interaction either in text or voice and so appear to be completely focused on my issue. The goal is to get a resolution more quickly for the customer and with less expense for the company.  Ram said in a recent video, "The future is having a real dialogue"  — one where conversations with computer systems are more than play a song, tell me the weather. Instead, conversations have turn-taking and artificial intelligence can execute judgment tasks like making suggestions, checking on insurance claims, and scheduling appointments. Perhaps even more important is that the AI knows if it needs to hand off the conversation to a human and transfers the background conversation in a way that makes that handoff natural.

Taking out the “Middlehuman”

Ram's background is "middleware" — the systems that get work done in our digital world. If we have a question, need to schedule an appointment, or would like beauty advice, we likely need to connect to systems that are talking with other systems. Our connection to these systems heretofore is based on keyboards and screens. We had to learn to create communications through keyboards, then keyboards and mice/trackpads, and now touch interfaces. Only recently have direct voice interfaces, powered by AI, become viable. I say direct voice interface because if I'm calling a phone number and talking to a person, that person is then likely engaging a keyboard and screen to connect with the systems that will let us reach our goals. What if we cut out the middle step and talked directly to the system, and that system reached out to the other systems to address my issue?

Top-Down and Bottom-Up

I tend to focus on how people achieve work targets by leveraging new combinations of our talent, technology, and technique, in a given time/context. My thesis is that to be effective in our increasingly digital lives and work, we all need to be systems designers and need to learn "think in 5T”™. For my lifetime, I expect this approach remains an important skill. That said, Avaamo is taking on the other end of the problem and getting the systems to adapt to us. To the extent that goals, intent, sentiment, and language can work via conversation, dealing with systems becomes easier for most people.

Rather than preparing people to design their own bottom-up applications of automation (versus top-down applications of automation to their work), some of the effort can be taken on by conversational systems such as Avaamo's; a top-down approach that makes our bottom-up work easier.

Next Steps and Thank You

We didn't have a chance to talk about whether or how to let a customer know when they are talking with an AI versus a human. I'm looking forward to learning more about that topic as Prof. Michael A. Santoro and I begin a project on the ethics of automation.

Thank you, Ram Menon, for the excellent conversation. I'm looking forward to seeing more of Avaamo's approach.