Building the first modern Conversational UI

It was an unseasonably warm summer day in SF when I revisited a tiny product feature from my past that would become one of the Web’s biggest changes in the past 10 years; SMS. I’m proud to say my last adventure, [digit], was the leader in influencing this change. But how did what I considered a tiny feature become so big and why are so many people duplicating it?

Hi, I’m Patrick and I’m a Product leader.

As a Product leader, one of the demons I carry is the knowledge that the Web is, in short, just a series of forms and databases. Several years ago, I’d reached a breaking point. I simply couldn’t design another valueless form. And those silly sign up/sign in/forgot password/FML loops? No thanks. At the time, it was a daily battle just to log into my own bank account, which as a young entrepreneur, is something I did more often than I’d like to admit. I could never remember which password worked for each account. All in all, I was annoyed and tired. And these annoyances prompted me to stay up all night making digit, the Company that fixed them.

But in order for digit to be a financial hero, we needed to dismember friction at many points. Checking accounts are more or less attacked with transactions 24/7 - funds are drained from you while you walk to work because of automated authorizations or hidden fees that go overlooked on your statements. You’re not informed of this activity and you’re left on your own to figure it out. I knew if we created another app that ran only when the User opened it, we would fail miserably. If they needed to be engaged all day, that was no better than logging into your bank every hour. It had to be working for you, at all times, swimming up-stream.

I recalled back to my first Company, where we used SMS to send movie recommendations. The engagement on it was incredibly high. Even as early as 2011, Users used SMS more than the Web UI. But could we seriously do finance through SMS? Is it too restrictive? Is this stupid? Would people get it? That’s just the short list of the questions we asked ourselves. But in my eyes, the opportunity outweighed the risks.

Demoing that prototype was funny. When people said, “OK, show me what you got!” and we opened a text message, we were met with a puzzled look. At that time, I couldn’t tell if they thought we were joking. One prospect investor said he liked our “little text” - we always chuckled at that. We weren’t expecting huge reactions but we knew something really big was ahead. It was like being automatically fed vitamins; except it was financial well-being and it was easy.

Pitching beta testers was even funnier. They simply didn’t trust us. “Wait so my checking account balance is going to be texted to me? For everyone to read!?” — Yep! We assured them that they couldn’t transfer money to other accounts than their own, so it was extremely safe. But at that time (2013) there was still a huge pushback in using SMS for something other than conversations with friends.

And whenever anyone asked if we were launching an iOS app, I just kept saying “Nope!”. If we made an app, we would dramatically increase the cost of doing business and lose focus on what made us great.

But guess what? 8 months in, it worked well. Really, really, well! Our users felt the convenience factor and value we provided. And things really started to go.

What’s extremely important to recognize is that our success had nothing to do with “the chat interface” or Conversational UI. In fact, we rarely focused on that. It was the last thing we worked on when we made updates to the Product. It was the brain behind digit that made it valuable and the fact that we used a platform our Customers were already familiar with versus forcing another app to keep top of mind.

All I need is a Conversational UI and a pitch deck ‘yo

I see conversational UI everywhere now. Most shockingly is that Facebook Messenger has made it part of its core API. It makes sense because the mechanism is quite easy to replicate and feel some value from.

But successful Conversational UI only works when the “app” is running parallel to your lifestyle— 24 hours a day, 7 days a week. It absolutely must do a complex job, do it at the perfect time in relation to your preferences and deliver that value automatically (without you needing to open “an app”).

Think about it, everyone forces Users to open an app in order to feel the Company’s value. How stupid is that? I have to remember to open this thing in order to appreciate it? Dumb… I wanted to install software onto my checking account and have it fight for me. That was my dream.

The Conversational UI is simply the software having an interaction with you about what it’s doing. It should never be about creation or management of relatively complex tasks. That’s best suited for the Web.

The future

I absolutely love the idea that software is something you can install into your lifestyle. Too many things require the User to interact with it. The machine should be smart enough to know how to deliver the value it offers - when to start and stop those jobs and much more. SMS is now incredibly powerful, it’s the general Consumers “terminal” — where multiple actions with multiple smart machines can all take place in one systematic place.

These machines are going to be tiny brains that we stick onto our lifestyles which contain real personalities, real traits and real cerebral power to be better than any human. But like any relationship, if it talks a good game but doesn’t have follow-through, you’re going to ditch it.

I’m excited for the future and I hope that my reflection helps you create an even better one for all of us.

Sometimes, carrying on, just carrying on, is the superhuman achievement. — Albert Camus


Help us by spreading the word