PeterBot — the chatbot which replaced my CV

Chat with Peter on Facebook Messenger

The idea

I would say the PeterBot project came from my common interest in artificial intelligence, but it was really hard at first to find a place for a bot in everyday life. Although Facebook, Slack and Telegram started to bring closer to people these simple applications called bots - especially since the F8 event in April — but I wanted something more than an other app which lets you order a pizza, so I started to think about how I can take this opportunity to the best use in my work, particularly in pitching for jobs. Then I started to look for resources, and find a great article from Esther Crawford, which opened my eyes about the right execution. (If you have time that’s where you should start as well)

What are these bots?


These tiny bots made of scripts are substantially alternatives to web and native mobile applications, and a lot of brands and developers are currently working on how they can integrate their existing product to these surfaces, so they can offer a closer brand experience for the audience. This integration eventually also means that in the future the number of mobile applications will be reduced, and everyone will strive for building the jack of all trades bot environment, which automates the lives of people, so you won’t have to open separate apps time by time, if you want to take an action.

The building up

I’ve chosen primarily the Facebook Messenger platform over the others because people usually keep their messages unlike on Slack or Telegram, where you can easily overlook the past messages and also because it seemed that resources are really well-documented. Not to mention the broadcasting option for follow-up messages, which make it easy to get in touch with the selected chat partner again after the first hype of talking with a bot.

But not so fast! tThe process started earlier: at first I had to analyse some of the existing bots like Poncho, CNN and others to know the limits of the surface and also mapping a lot of existing solutions. Actually after hardly trying to code elements using GitHub forking, Heroku, Smooch and asking questions on Stackoverflow (which was all great btw), I found a great solution in Chatfuel, so the creation of the MVP from existing modules took only a couple of days.

The biggest challenge was creating an engaging conversation in a short period of time, since the attention span is really low. I started with a quick card sorting to make a selection of certain topics in a typical conversation. Then using all the different topics, I made a shortlist of them around 3 great storylines: early years in my life, education and career. After doing a little sitemap-like model of the different conversations I could say at this point by 90% what will be the actual user flow of my PeterBot application.

I’m still in the beginning of this journey but I would say I wasn’t wrong with the estimation because after the release of the bot the frequency moved around only maximum 3–4 questions / person, and they rarely asked questions aside, so structuring properly the information can work in a human-bot communication so far.

The future of bots

The future can be really versatile in terms of using these bots, but it seems that online customer service, community management and several online-based jobs can be successfully replaced with structured conversations. One of the greatest challenges now is actually finding these bots, although I’m happy that there are some initiatives on this, like the Botlist, or Botwiki communities. In case of my bot, I would say that bots like PeterBot are really useful assets as a CV during the application, because it’s also beneficial from the employers’ side who can quickly and conveniently chat through the basic information about the possible participant and a great talk about the bot can break the ice at any time.

At last but not least, I have received very good feedbacks in relation to the conversations, so it seems that trying to design well the user experience can pay off. If you haven’t tried my bot yet, you should do it right away!

If you want to create your own bot I can recommend you numerous resources, and the amazing Esther Crawford who was the pioneer of creating bots like mine, and also don’t be shy to tweet me if you have a question or recommendation.

// Please tap or click “♥︎” to help to promote this piece to others. //