Poor misunderstood bot…

I’ve been in involved in technology for over 20 years now – and without fail there will be a new flavor of the year. When Facebook launched their bot for messenger platform – the spotlight shifted, and everyone’s rushing to this next goldmine (happened with “web 2.0”, happened with “mobile”, happening with the internet of things, and happening with big data).

Here is a chart of what is probably on everyone’s mind right now in terms of approaching the bot opportunity:


It’s that simple… Or is it?

Unfortunately, if it were – it would probably be equally simple to earn $10,000 an hour working from home with no work experience (btw if anyone has ever followed through on one of those ads please call me).

Now, after studying it through – the reality is a lot more like this:


As with all trendy buzzwords that pop up like hipster coffee joints in old shop houses – you have to start delving into the meat in your business approach to it.

How do I start demystifying?

This isn’t going to be a tutorial on bots. There are other much more competent sources for example this article by Matt Schlicht among many others. But I just wanted to point out how to frame your thinking when applying chatbots to a business (or even as a startup) – and in particular, focus on how not to think of the chatbot.

Myth Number 1 – Build a bot and they will come beating a path to your door

Nope. Not going to happen. You see, a bot is just a channel – a communication channel. In that context, it is no different in how you treat it just as you think about your website, your search presence, even your TV communications.

It is just a different channel by which your customers can reach you

That’s about it really. Of course, whole theses could be written on how different a channel messaging really is – but by and large this is how you have to frame your thinking on bots. And conveniently, this segues into the second myth:

Myth Number 2 – Bots are the silver bullet

Remember – a bot without a problem is like a pen without ink. You have to think what problem you want to solve. At the heart of the matter, there must be a problem to solve. If there isn’t – a bot is nothing but a novelty.

And a useful way to think about it is to think about a problem you need to solve – and more importantly why using messaging as a platform contributes to the solution. Here are some examples.

  1. Reducing customer support costs – because people frequently use messaging to get help.
  2. Automating smart home tasks – because it is much easier to text “switch light off” then to load up a website, navigate and then click a button.

Myth Number 3 – Bots, Artificial Intelligence, Machine Learning, NLP are the same things

This will be really obvious to the technical folks – but many business people forget or don’t even know this. As such, a common problem is that when discussing bots, these concepts tend to get conflated – which is a dangerous waste of time in meetings when misunderstandings occur.

To help, I’ve tried to make the following distinctions:

Bot – An automated virtual robot that replies conversations on your behalf (you should know this already!)

Artificial Intelligence – The brains behind the virtual robot (this term extends beyond this landscape though e.g. gaming). This is how the bot thinks. This is done through various ways – but you’d be surprised how rule-based most of the current solutions are (go read up on AI Markup Language).

Machine Learning – This has nothing to do with AI. Rather, I like to think of it as a separate subject matter altogether. This concerns more of giving an algorithm or library a training set (e.g. images of flowers) and training it to recognize patterns and probabilities. This article is very helpful to understanding it in that context.

Natural Language Processing – How to parse language into something that computers can more easily comprehend. Don’t underestimate how complicated language can be – and NLP is all about simplifying language into pieces and making sense of that.

That’s Three!

And hopefully that would give you some starting points on how to shape your bot strategy for your startup or business. Obviously there are heck of a lot more detail and meat around this but for starters – this post serves to frame thoughts and not really throw you into the deep end.

If you are ready to go into the deep end – feel free to send drop me a mail at adrianl@brazn.co!