Thought Leaders #4: Braden Ream, CEO of Voiceflow
Meet Braden Ream, CEO of Voiceflow. He currently resides in San Francisco, but grew up in Toronto, Canada. Since the founding of the company in 2018, it has raised a total of 25M of public investment and the platform is used by 120,000 people worldwide.
He doesn’t really need an introduction, since anyone who’s been around in our space for a while knows who he is. I personally know him as a guy who loves his own product and never shies away from doing a live demo himself. I had the pleasure of interviewing him, so we could talk about the importance of conversation design and the future of AI assistants.
Let’s dive right in!
So, Voiceflow hosted its first LLM Hackathon PromptHacks recently. Was it a success?
Yes! It was a great success. We had over 400 registrants and 63 projects created over the weekend. We’re so happy that people got to play around with all the new LLM features.
We want Conversation Designers to be able to play with these new technologies now, and get a feel for where the space is going. We believe the best way to learn is to dig in! Sometimes folks just need a little nudge, and so we hope PromptHacks was that nudge for people to learn about how CxD is going to change (for the better) over the next few years.
You started Voiceflow in 2018 and you’ve been growing your company ever since. The company has also proven to be quick to adapt to the advent of large language models. Can you talk a little bit about your journey?
It’s been an exciting one, and we owe it all to the support of our customers and community. We started the company in 2018, but originally as a product called “Storyflow”, where we were building interactive Alexa Skills.
Our team was 4 people back then. We found the collaboration process between our Conversation Designers (myself and my co founder Mike), and our Developers (my co founders Andrew and Tyler) was incredibly difficult and the product grew out of our own collaboration problem. We needed a better way for Conversation Designers and Developers to work together, and we couldn’t find a tool that met the need so we built Voiceflow!
I think we ship faster than others despite our team being half (or more) the size of other bot development platforms because we invest so heavily in infrastructure. We’ve spent millions of dollars on our CI infrastructure, cloud ops, and internal tooling, so that developers at Voiceflow can build fast while maintaining quality. We recently doubled the size of our engineering team (again), and these people are all still ramping. If you think we ship fast now… just wait until 3-4 months from now.
What made you want to empower the conversation designer in the first place?
I don’t think traditional platforms cater to the Conversation Designer as much, because they are seen as a lesser, secondary audience to Developers. At Voiceflow, however, we believe that the best conversational AI teams view design and development as equal counterparts when building AI assistants. Conversational AI hasn’t caught up to this viewpoint yet, so we’re doing our best to empower Conversation Designers by giving them the tools they need to get their seat at the table.
To be fair, it’s part of how things naturally develop: design is initially undervalued when a new interface or platform launches — the same was true for mobile and web. The value when these new platforms launch can often be unlocked mostly by development as simply having a mobile app or website is novel, and consumers don’t have benchmarks for what “good” looks and feels like.
It’s only as competition within a medium or platform arises, like everyone having a website, that design begins to matter. Now that AI assistants are everywhere, the need for great Conversation Design is growing in importance as consumers better understand what good conversational experiences look and feel like, and detest poor bot experiences.
We all know and share this feeling that the conversational AI space is moving at breakneck speed. How do you keep up?
For me, it’s all about playing with the latest technologies as early as possible, and keeping up with our own Machine Learning team to discuss the latest papers and advancements coming out.
My advice is always to get out there and play with the latest! It can feel scary at times how fast things are advancing, but the only way to conquer fear of the unknown is to start working with these new tools and understanding both their benefits and limitations.
Finally, let’s talk about the future. LLMs are changing the way we build AI assistants. How is it going to affect Conversation Designers – now and in the future?
So, first of all, personas are going to become increasingly important. Think of it like this: if you were to train a human how to talk with customers, you would give them examples of what to say, but largely you’d focus on tone and personality. You wouldn’t have a human read directly off a script — so why should we have bots do this if they didn’t have to?
Secondly, LLMs are incredibly powerful, but won’t be used in the enterprise as immediately as the tech demos are available. There’s a lot of tooling that needs to be built first. As LLMs have a tendency to hallucinate, meaning give false answers unknowingly, it’s unlikely they are used for mission-critical enterprise use cases as of right now. I suspect LLMs will be used for less critical use cases in the interim, like being able to answer Q&A etc.
Lastly, Conversation Design isn’t going anywhere. In fact, I think it’s going to grow exponentially as a career path. I think what LLMs will do is explode the number of use cases where AI assistants can be applied and what tasks it can handle, growing the number of assistants in the world exponentially. It’s like air travel: as the cost of flights came down, people traveled more, and the number of pilots grew. As AI assistants will become more ubiquitous, the need for conversation designers will grow with it.
Follow Braden Ream on LinkedIn
This interview is part of the Thought Leaders series: a collection of written interviews with industry leaders and experts in conversational AI. Feel free to reach out if you think I should talk to you.