• BURN MEDIA
    • Motorburn
      Because cars are gadgets
    • Gearburn
      Incisive reviews for the gadget obsessed
    • Ventureburn
      Startup news for emerging markets
    • Jobsburn
      Digital industry jobs for the anti 9 to 5!

Facebook Messenger’s bots are mostly shit, but they’ll get better

Bots are the new cloud, it has to be said. They’re the latest trendy thing in the world of tech, as the likes of Google and Microsoft get in on the action over the past few months.

Another name to add to the list is Facebook, actually beating its rivals to the punch by launching an army of bots worldwide (for the most part). In fact, Facebook claims that there are 11 000 bots available for its Messenger platform at present.

Unfortunately, its bot efforts, while still in the early stages, are rather poor right now.

It’s like a robot phone call (and not in a good way)

The biggest selling point of the new generation of bots is that they’re supposed to be smarter and learn from you over time.

But most of Facebook’s bots are fundamentally identical to bots of yore (such as MSN Messenger and Mxit), featuring a standard conversation tree system that don’t understand contextual, natural language very well — if at all. It then becomes a case of figuring out the limited number of commands at your disposal (or you get a command list, if you’re lucky).

Related: Are smart bots the way forward? Microsoft says yes

It’s a far cry from rival bots, as Microsoft claims its bots will use machine learning and a host of other techniques to gradually become smarter and understand natural language.

Thankfully for Facebook, it isn’t standing still on this front. The company recently announced that bot developers will get access to the engine behind Facebook’s M assistant. Will this make for better bots than rivals though?

Microsoft and Google have a lot more invested in natural language processing and machine learning than Facebook, owing to their Cortana and Google Now products. So if anything, we can expect their bots to be a step above Facebook’s current crop. Here’s hoping Facebook surprises us.

Discoverability? Hahahahaha

What’s the point of having 11 000 bots on your platform if there’s no easy way to view them?

Instead of having a centralised listing (or phone book, if you will) of bots, Messenger requires you to search for bots in the “search” bar within the app. In other words, there’s no native way of discovering interesting bots.

Related: Chat bots, drones, AI: Facebook’s grand plan for the next decade

Fortunately, there are third-party websites serving up listings, such as botlist.co and Messenger Chat Bots. But when Skype has a directory of bots, and its bot functionality is still in preview, it’s pretty inexcusable.

Nonetheless, this is another department Facebook needs to tackle if it wants its bots to be a major success (or a success in general).

Bots you should try

Despite these challenges faced by Facebook, there are still a few Messenger bots worth checking out.

For one, the Digg and TechCrunch bots are great for tech news junkies, serving as a substitute for newsletters and dedicated apps in a sense.

These apps are able to give you recommended news stories, or you can search for stories by tag. A rather convenient system, then.

In the mood for something more interactive? Then you’ll be glad to know that classic text adventure Zork is also available on the platform. Granted, you’ll encounter quite a few “I don’t know that word” moments, but it’s still an interesting use for bots.

Need to make something for supper but out of ideas? Well, the Dinner Ideas bot will send ideas to you daily or once a week. If that’s not enough, you can also submit a list of ingredients and it’ll spit out recipes.

Author | Hadlee Simons

Hadlee Simons
Terrible puns make Hadlee Simons difficult to work with, but he brings over seven years of tech journalism experience to the table. When he's not at work or watching motorsport, he's in the foetal position on a jiu jitsu mat. More
  • Ilya Geller

    Yes, Facebook could improve its bots. How?
    Facebook must follow Laws of Nature. For instance Facebook should use natural parsing, which Google uses – see http://www (dot) wsj (dot) com/articles/googles-open-source-parsey-mcparseface-helps-machines-understand-english-1463088180 – and calculate internal to language statistics, as Google started at Nov 12, 2009 Google already calculated “significances” of a keywords – https://www (dot) webmasterworld (dot) com/google/4023900.htm. For instance, there are two sentences:

    – Alice goes up.
    – Alice goes left, Alice goes right.

    Evidently, that the ‘Alice’ has different importance into both sentences, in regard to extra information in both. This distinction is reflected as the phrases, which contain ‘Alice’, weights: the first has 1, the second – 0.5; the greater weight signifies stronger emotional ‘acuteness’; where the weight refers to the frequency that a phrase occurs in relation to other phrases.