My chatbot is one of the thousands of bots living in Facebook Messenger. I created it to keep interested people up to date with blogs I write and to help new audiences find my content. The idea was to use chatbot like newsletter while, of course, taking advantage of all capabilities this new medium offers.
So now I have a chatbot that educates people about live chat, sales chat, and chatbots. Fun.
You can chat with my bot here: m.me/ilkkachats
Why did I do this
As said, I needed a channel I could use to reach readers of my blog. A newsletter was my original plan for reaching my audience but in the end using email to share content about chatbots just felt wrong. Using a chatbot just felt more natural.
I am a big believer in chatbots. I believe all kinds of virtual assistants and chatbots are the way to go. Instead of setting up and going through a steep learning curve about tips and tricks regarding email marketing I decided it is more fun and useful to jump on the new medium and gain more understanding about chatbots.
Also, I am a firm believer in social selling. I am not going to hide the fact that my chatbot is a part of my social selling efforts. If I am right, and I can build a bot that is both fun and educating (no pressure here) to use, my chatbot opens a whole new way for me to interact with people sharing my interests. So, one of the next things my chatbot is going to learn is how to connect human-me into the conversation.
Go and chat with my bot to learn about live chat and chatbots.
How did I do this
I used a free chatbot platform called Chatfuel. My main reasons for selecting this platform were 1) no coding required, 2) it is free. To be honest, I did not do much research for the platform, I just ended up tinkering with it, and here we are.
Was creating a chatbot difficult? No, I wouldn’t say so.
It took me maybe an hour to understand how Chatfuel’s platform works and digest the logic for building a simple chatbot on it. Timewise from the moment I started to the point I “released” my bot into the wild, it took me a weekend and few evenings to set it up. Maybe around 15–20 hours altogether for most of the capabilities to be up and running. After that, I spent additional few hours here and there to update content and conversation flow. I estimate this time being around same or even less if I had set up a newsletter (templates, settings, signups, etc.) from scratch and optimising those.
What did I learn
During those 20+ hours, I build two chatbots. The first one was techwise ready to go but I though there was absolutely no reason for anyone to use it. The latter one (the one available now) is built on that “template” with more focus on narrative. Focusing more on what the user might need from the chatbot instead of what I wanted from the user (read my blogs, pretty please). It is not perfect now, but it is way better than the first one.
Even including obvious lessons about using the platform and learning the insides of a chatbot, the focus and effort required for the narrative were the main lesson for me.
This exercise also taught me the power this new medium has. Creating a chatbot is surprisingly easy. I would say if you can set up a newsletter, you can have a (simple) chatbot. As this is something, we can expect to get easier and easier as bot platforms evolve. Besides, I expect there to much more two-way interactions initiated by a bot than by a newsletter.
I am going to go on the record here and say that chatbots are the equivalent of newsletters in the future.
What am I doing next
My chatbot is not ready. I don’t think there is such thing as finished chatbot. There is always something you can tinker, optimise, add, or teach to it. While creating this somewhat simple version of digital me, I come up with a long backlog of features I want my chatbot to “learn”. Top of my list are the ability to transfer conversations to a (human) me and collect feedback and ideas for better content and improved bot. Also, I am expanding its conversational capabilities to answering simple questions regarding live chat, sales chat, and chatbots instead of just recommending (my) content.
Also, what comes to long term plans, bots name might contain a hint. Like the name of the chatbot, Sales Fanatic might indicate, my chatbot probably won’t be forever focused only on topics above.
So you made it this far. What if you would take my bot for a spin?
This blog is originally published on Medium on May 22, 2017. Read the original.