The Battle of Virtual Personal Assistants
In one of those “why didn’t I think of that!?” moments, Facebook is rolling out M, the personal assistant who lives in Messenger. It’s basically like any one of your friends, except when you ask for advice it will be like Yelp on steroids.
It’s a really clever idea for both users and Facebook as a business. Here’s why:
1.
How do you actually feel speaking with Siri and Cortana? Still pretty natural, but I doubt many would think of either assistant as a friend. In a huge twist of irony, I’d argue the downfall of voice-based assistants is that you can hear their voice. They have an identity, and it isn’t one you know. With M for Messenger, the genius is a combination of context and delivery. You can’t hear M (for now). It’s left to your imagination. And given M appears alonside a crush of conversations with your IRL friends, it’s easy to think most people will be a lot more comfortable, whether consciously or subsconsciously, asking M for advice instead.
2.
Beyond those tricky feels, let’s talk about actual usefulness.
Siri, Cortana, and Google Now are awkward. There, I said it.
Yes, the novelty of it is cool, but we’ve still got a long way to go before voice-based assistants are actually the most efficient way of going about things. Fun? Yes. Most effective? No. It’s not yet consistently smooth UX. While it’s helpful for hands-free activities, you and I have both seen many people using it when their hands were perfectly available. The reason for this is surely part gimmick (“Ha, check this out!”), but partially because there aren’t any well-known alternatives.
With M, Facebook just offered one, and it’s text-based. (No, your Google/Yelp search doesn’t count, that’s one too many steps to get to. More on this in point 3.)
Just like we will always look a little insane talking to ourselves with our bluetooth earpieces (to say nothing of Google Glass, RIP), we will also always prefer to not make a scene and discretely shoot a quick text instead. The reason texting has taken off as the preferred mode of communication is that it’s socially versatile. Logically, a text-based assistant will also be more versatile. It’s a clever niche between chatting and a Google search.
3.
Your phone isn’t for phone calls anymore (yes, I know, cue rant about modern society). It’s for text-based chat, perusing what’s trending, and searching for local advice. Facebook has the first two covered, but not the third. Which means you have to leave their app and resort to Google search, a virtual assistant (Siri, Cortana, etc.)… or circle back and ask a friend.
Ask a friend.
I can only imagine the eureka moment went something like this:
“(Insert Name) made some really great suggestions when I was in Chicago last week. But I don’t know anyone in Boston, wish I did…”
“Yeah, they should make something like that with AI…”
“…………”
Facebook have one of — if not the — largest databases of recommendations, between what its users share and reviews on its increasingly popular company pages. M is probably the best use of it to date, which in hindsight seems like a no-brainer, because with M, Facebook now has the previously mentioned trifecta figured out. Go to Facebook to chat, see what’s trending, and receive great recommendations.
4.
M is going to have actual people fine-tuning/ augmenting the recommendations.
Via Rob Daniel, Product Manager at FB:
Friends! Today we’re rolling out M, a personal digital assistant inside of Messenger that completes tasks and finds information on your behalf. It’s powered by artificial intelligence that’s trained and supervised by people
I am fairly convinced this will be something we won’t think is making a big difference in our M experience, but it will. Just like many people still love Pandora because the recommendations are people-powered, or loved Songza for its mood-based playlists, the value of the human touch shouldn’t be underestimated. When people are searching for a cool place to eat, or a great shop, they are looking for something special. Sometimes — often, actually — that’s something out of left field that doesn’t add up in a typical algorithm.
“It can perform tasks that none of the others can,”
— David Marcus, VP Messaging Products @ FB
So: take Facebook’s massive database, mash it with a surely worldclass AI, and then add on a team of humans to spice things up? That might be the secret sauce that really takes it over the top. (Yes, again a hefty dose of irony when talking about an AI assistant.)
I think it would be especially cool if Facebook opened this up to its users to augment directly as well, and also integrated it into Mentions (their app for celebrities/influencers). Imagine asking for where to grab a slice of Pizza, and seeing that Joe and Tina explicitly recommended X and Y as one of their Top 10 favourites, and that Dwayne “The Rock” Johnson recommends Z. Nifty tip.
5.
If you’ve got $FB stock, you should probably be cautiously excited. I would be surprised if this doesn’t get integrated into their advertising platform, with M becoming the most sought-after/expensive real estate (and not just on FB, but of all online search). In the same way Google and Yelp provide recommendations based on what you search, M could easily do the same. Again, however, because it’s an immediate search, rather than that extra delay to pop into Yelp or Google or open an assistant, I predict that will hold tremendous value. The first answers people get usually become their reference point, and more often than not what they end up going with (if the recommendation is decent). Would this take away from the whole I-could-just-as-easily-be-texting-with-a-friend-right-now! psychological aspect? Maybe, but not necessarily.
“We start capturing all of your intent for the things you want to do,” says Marcus. “Intent often leads to buying something, or to a transaction, and that’s an opportunity for us to [make money] over time.”
If M can provide a more efficient service than its competitors, Facebook can boost the number of people using it on mobile, and eventually spur revenue from their transactions. That’s the kind of win-win Marcus was brought in to accomplish at Facebook, which in June 2014 hired him away from PayPal, where he had been CEO.
So that’s my two cents on M. I’m pretty excited to give it a spin. I think it might go on to be one of Facebook’s most transformational new releases in years. While it’s easy to claim many of Facebook’s other “innovations” are imitations of other trending products, I see this one as a genuinely clever release. It’s subtle, but it’s significant. Let’s see how it pans out.
