For no good reason besides satisfying my own curiosity, I made a Twitter bot. It’s name is FundraisingBot and you can follow it on Twitter here.
Over the past year I began noticing a surge in articles about bots, like “Bots, the Next Frontier” in The Economist, and “Forget Apps, Now the Bots Take Over” in TechCrunch. Each hypes the opportunity that bots offer, mostly in the area of automating personal assistant-like tasks. Think ordering dinner, sending reminders, booking a hotel room, etc.
If you still have no idea what I mean when I say ‘bot’, think Siri. Not a physical robot. These are programs that automate often simple routine tasks. The best ones interact with you in a human-like fashion, replacing the need for a human to carry out these simple tasks.
Facebook and Microsoft have built platforms for developers to create bots, and over 45,000 developers are currently putting them to use. The most enthusiastic bot evangelists believe the technology has the ability to open up entirely new areas of business and commerce. The more pragmatic of us see the value in automating some tasks, freeing up humans to do what they do best – build relationships, solve complex problems, and empathise.
So why did I make a bot?
I want to be completely up front about one thing – I am not a programmer. I can make minor tweaks to the most basic html code, but that’s it. To be more specific, I made a bot account, not an actual bot.
Why make it at all? Because I wanted to see what all the fuss was about, and to find out if there might be any practical application of this technology for a charity. And if so, can someone with no coding abilities set one up on their own?
Behind @FundraiserBot there are actually two applications running simultaneously: one that generates random, non-sequitur posts which are a mash-up of other fundraising accounts’ tweets; and a one that can, to a limited degree, answer questions that meet a criteria I’ve defined, with answers that I’ve prepared.
Function #1: Random, non-sequitur Tweets.
This all started for me with this article called “How to Make a Twitter Bot in Under an Hour” that popped up in my Medium feed. I thought that sounded fun, so I went for it. I’ve also noticed that many of the accounts (personal, and corporate) that tweet about fundraising start to blend together. Similar blog titles, similar advice, etc.
This absolutely took me more than an hour. I had to download some free programs, troubleshoot some errors, and take some much needed breaks. Lauren Orsini’s Five Steps To Build Your Own Random Non-Sequitur Twitter Bot turned out to be the more practical how-to that helped me get this up and running.
What it does is simple. I’ve supplied a list of fundraising-related twitter accounts, which then get mashed up into new tweets. The script is available for free, and does the heavy lifting for you. It does it’s best to ensure the tweets make some sort of sense, and are sent on a schedule.
Here’s a sampling of what this looks like:
Function #2: Conversational tweets and link sharing.
Randomly generated tweets can be good for an occasional laugh, but gets old quickly. So up next was to try to figure out how to help @fundraisingbot converse with others.
When I read that Microsoft’s bot framework is available for free, I thought that was my simple solution forward. At first glance, it looked like I’d be able to easily add some some artificial intelligence (AI) capability to the bot which might add some fun, and hopefully useful, layers to the bot.
I struggled for days trying to get the AI bot running before running out of patience entirely. I’ll likely come back to it sometime and give it another shot. Until then, I searched for a plan B. That turned out to be Chatterbot, which again, frustrated me to no end, and ultimately never went live. The problem being, the instructions are written by programmers, for programmers. There are prerequisites that at this time, I don’t have. Take a look at the instructions, and you’ll likely know how I felt.
Then, on to Plan C. This turned out to be a site (or more specifically, an app) called Cheap Bots Done Quick. This free tool allows you to use a language called Tracery. There’s a learning curve to the language, but do-able. Within an hour I was comfortably writing within it and troubleshooting errors successfully. Once running, it can send tweets on a schedule I set, from the pool of tweets and interchangeable terms I supply it with. It is not artificially intelligent in any way, but it is still a bot. I follow others manually, but I never interfere with the tweets. If someone asks a question and the bot doesn’t respond, I’ll adjust the code so that in the future the same question should trigger an answer.
You can see the full code here if you’re interested. But I think it would be much more interesting to just go ahead and ask it some questions instead.
Where do we go from here?
I’ve only set up the most rudimentary bot as far as Twitter bots go. No AI or learning capability, and finicky conversation. Bots aren’t limited to Twitter either by any means. There are Facebook Messenger bots, Skype, Slack, Reddit and more. I’ve only started with Twitter because it’s where I’m most comfortable.
I’d like to see @FundraisingBot become a resource to students and professionals, referring them to resources and providing quick, easy tips on a regular basis.
For charities, I’m still undecided if there’s a practical use for bots beyond a clever campaign, or simple redirects. Charities, and especially fundraisers, rely on the honesty and authenticity of their relationships. To outsource almost any component of relationship building could be short-sighted.
Charity: water takes donations via Facebook Messenger bot via UK Fundraising
These 2 Brands Made a Facebook Messenger Bot to Draw Awareness to Ethiopia’s Water Crisis via AdWeek
Bots, explained. What’s the point? Or the business model? via ReCode
Get Ready For The Chat Bot Revolution via Forbes