Why I Think Digital “Assistant” is the Wrong Metaphor

• ~800 words • 4 minute read

I have a theory about digital assistants. I won’t pretend it’s backed by any real research or data, but it feels like a theory with legs. There is a long lead-in before I get to digital assistants and I don’t quite stick the landing, but bear with me. It’s an idea I want to throw out there.

First, the obligatory backstory: Smartphones have changed plenty over the past decade or so. Many of us walk around with instant communication devices in our pockets all day. We’re seldom more than a glance, tap or whisper away from looking up all kinds of information, translating our thoughts to another language, checking the weather forecast or getting directions to where we need to go.

That last one is key to me. All these years later I still think that having a map in your pocket is the single best thing about these digital devices. It’s not even close. Email in my pocket was a novelty, at first, and ow the bane of my existence.

Last week I was in New York giving a presentation at BrooklynJS and had no trouble figuring out which trains to take as I scurried about the subways. It was the first time I’d really explored the city since I’d visited when I was 20, well before the advent of smartphones. I can remember navigating the subways then and walking literally hundreds of blocks in the sweltering summer, somehow without issue, but I’d be lying to you if I said I could remember exactly how I figured it all out.

Here’s the thing: I think maps were the original — and perhaps only — “killer” app for the simple reason that the benefit was incredibly obvious. It’s so obvious it’s kind of brilliant. We all have to navigate from point A to B at some point and having an accurate, consistently updated picture of the place we live in our pockets makes it almost mindless.

The other thing that made it so effective is you didn’t really have to teach people how to it. It wasn’t a major paradigm shift. We took maps, something human beings have used for centuries, and made them even better.

That leads me, finally, to digital assistants like Siri, Alexa, Cortana, whatever Google’s is called, if it even has a name. I think that having a generic “assistant” is well outside the day-to-day experiences of most people. I’ve never had a person in my life dedicated to remembering my meetings, taking note of my tasks or generally running errands for me.

Reading assessments of how well a particular digital assistant performs relative to others is also an interesting experience because it seems everyone has wildly different expectations of what they should be able to do, let alone how you should talk to them.

I don’t think digital assistant technology has changed the way we live as profoundly as maps. That’s not because AI and machine learning are still somewhat nascent. That’s because, honestly, having an assistant for everything in your life is some bourgeois, first-world bullshit. It’s a feature built for CEOs and the 1% accustomed to servants and butlers — not an experience 99% of us grew up living with. I think most of us lack the context for how to use a assistant.

So here’s what I think: I think the entire digital assistant industry needs to be rethought. We need this thing to take on other roles, beyond the generic title and unclear expectations surrounding the title “assistant.”

If you tried to sell me on the idea of a digital teacher or mentor, then I think you’re positioning the technology in a way that’s more universally relatable. An AI that can teach me how to play the piano or give me a lecture about the history of computers while responding real-time to my questions is much more interesting to me. Teachers and mentors are things I’ve had in my life and something I have a more immediate understanding of how to interface with.

I realize what I’m describing in someways are specialized “bots” that focus on specific things instead of trying to be everything to everybody. There’s a parallel in my mind between building software with a microservices architecture instead of a monolithic one, the “serverless” revolution we’re undergoing in web development right now and rising popularity of functional programming.

I’ll leave it at this: as it stands today, I find most digital assistants come off as gimmicks and feel the “assistant” label makes the technology unrelatable. In shifting the metaphor from “assistant” to something more specific and universal — teacher, mentor, tax advisor, personal trainer, counselor etc. — I think there’s a chance to build something that might more profoundly touch and change the way people interact with these devices.