Doing the ‘Robot’

The robots are coming. No. The robots are here. In case you haven’t noticed, almost everything we do online includes some form of bot intelligence. Facebook, Google and Apple are investing heavily in machine learning (the technique where we teach machines to teach themselves). But what’s all the fuss?

These companies have managed to outsource huge amounts of work to automated algorithms, or ‘bots’. When you upload a photo to Facebook, do you think there is a CIA-trained face expert on the other end, trying to figure out which of your friends are in it? No, like many repetitive tasks, this is now the work of bots. And they are being used more and more.

Bots will take over many jobs we consider to be uniquely human – via CGPGrey

But why do Facebook and Google get to have all the fun? Well they don’t, you can get involved too.

How can you get involved?

Lets start at the beginning.

You want to use bots for one reason. Saving time.

The best way to do this is to take the work you do on a daily basis and hand it over to a bot. But you probably think “Thats impossible, my job is so uniquely human, it expresses my creativity and my very soul. No robot could ever do that.”

Wrong! Bots can help with an extraordinary array of tasks and are already being used in technical practices such as medicine and law. Let’s look at a few examples of how bots can be used effectively.

Level 1: If day of week is Monday, then schedule a new Twitter post.

The most humble bot is the rule-based bot. Given a simple input (day of the week), it can choose between a number of actions to perform (post to twitter, alert a colleague, download a new report). But even these humble bots (aka, IFTTT, If This Then That) can dramatically reduce your time spent on repetitive tasks. Here are a few examples:

If This Then That. A fundamental operation of computer programming.

If This Then That. A fundamental operation of computer programming.


Marketing automation. Save time by scheduling posts and automating customer emails. This can save a huge amount of time spent on your inbox, reading and writing the same emails over and over again. Bots are faster, more accurate and work 24/7. Try Autopilot or Buffer.

Internet automation. Automate a near limitless number of tasks by linking data (weather or news events) or apps (MailChimp or Skype) to custom rules you create yourself. Try IFTT or Zapier.

Data automation: Time to get your hands dirty. The simplest type of data automation is the IF function in Excel. Instead of manually editing/renaming/moving your data, write functions to do it for you in a fraction of the time.

Level 2: Smarter Automation

The benefit of rule-based bots is that they can be applied almost anywhere (see Smarter bots, on the other hand, are often limited to the domain they were built for.

Messaging Bots. Advances in natural language processing have allowed us to communicate with bots in plain english rather than using cumbersome software interfaces. Talk directly to SlackBot (Slack’s in-app assistant) to quickly create reminders or outsource work to other team members. Or hire Amy, an email chat bot that will schedule meetings for you, saving the back and forth hassle.

Advertising Automation. There is way too much data in social media for any one human to process, let alone make informed decisions about. Why not outsource decisions about where to advertise and who to advertise to at each second for the day to a bot trained for that exact purpose. Try Needls.

Gmail SmartReply. Based on a huge amount of email data, Google has built an algorithm to make helpful suggestions in response to emails. Use it to quickly send thoughtful responses with a single click.

Level 3: Custom Bots

While the tools mentioned above can save you time on the more repetitive tasks, there is a limit. A lot of these services can only be used for specific purposes.

Netflix has a complex algorithm to provide suggestions to its movie-hungry audience (known as a recommendation engine), which they value at $1 Billion per year. Google and Facebook have intelligent classification algorithms that can segment and understand their users with frightening detail. And logistics companies are using bots instead of drivers to optimise delivery routes.

Machine Learning. Bots teaching bots.

Machine Learning. Bots teaching bots.

The intelligence behind these services are too specialised to be simply integrated into your own workflows (at least for now). To offer really valuable machine intelligence you will need data scientists who can tailor the bots to your specific task (looking for one, start here). So why not create your own bot to do your marketing emails for you, recommend products for you, target useful customer segments for you, or chat with new customers for you.

Finally, once you have set up your intelligent bot minions. Hit START and take a break, trying not to think too much about how the bots will one day take over the world.

Leave a Reply