Corona is spreading…and so are the bots

Now that Corona has large parts of the world firmly in its grip, more and more Corona chatbots are appearing. While these attempts are in itself laudible, effective use of chatbots as crisis communication tools might benefit from less bots, not more. Instead, focus on a solid content strategy, sustained maintenance and clear ownership.

Photo by Joshua Rawson-Harris on Unsplash

Whereas Google may have banned Corona related voice actions and apps from its channels, this is clearly not the case for Corona related chatbots. This morning, around 15 minutes of surfing the Web, Facebook and LinkedIn resulted in around 50 coronabots, in different languages, with different purposes and in different states of maturity.

Don’t be part of the problem…

When a crisis like Corona appears, there is a natural tendency for sharing information. Unfortunately, not every piece of information is accurate. Apart from the obvious scams and people that try to monetize on the Coronavirus by offering fake cures and treatments, information may simply be outdated, unclear or not applicable to your situation.

Adding more chatbots in this case, doesn’t add to a solution: how do people know which chatbot provides the most reliable information? Where the information comes from? Whether it’s applicable and up-to-date?

…be the solution: help the officials

So now what? To be honest, I think the best thing we as a conversational community can do, is to reach out to the official government bodies, offer them our help and insights, and make sure that we enable them to take control in providing reliable, accurate and timely information to the general audience. Through chatbots, in addition to their other channels.

Once that link is established, yes, we do need the best conversation designers, the brightest technology partners. But until then, it should be official bodies that should be in control. Why? Because they:

  • Already have a content strategy in place. They carefully consider which information to release, on which channels and in which amount of detail.
  • Own the source content. They check content for accuracy, completeness, relevance and timeliness.
  • Are known as reliable sources. When I can choose from 50 chatbots, I go for the one that’s got Reliable Source stamped all over it.
  • Have processes in place to maintain content. This for me, is the most important aspect of any content project, and is paramount in crisis communication. Outdated information is misleading, and in the case of Corona, where every day sees new recommendations, can be dangerous.

Sure, but I’m a bot builder with spare time…

If you do decide to build your own Corona chatbot, here are some tips that help you build a reliable and trustworthy chatbot.

Photo by Ian Schneider on Unsplash

Stay close to the source

If possible, automatically retrieve information from a trusted source.

If you do craft your bot responses by hand, make sure that you stick close to the source text. Any kind of interpretation, specification, shortening or making a text more conversational runs the inherent risk of altering some of the original meaning. That’s a responsibility that’s not to be taken lightly.

Top 4 Bot Tutorials

1. AWS setup for Deep Learning

2. How to Integrate API.AI with Chatfuel

3. Build a working SMS chat-bot in 10 minutes

4. Chatbot Conference 2020

Date and source your information

Within your bot, clearly state where your information comes from and the date of your last update. If you use different sources, make sure that they don’t contradict each other. In such cases, it may be worth mentioning the source in each individual response.

Check what’s already there

Before you start building your own bot from scratch, check whether there are people working on the same project in the same country or area. Besides sparing yourself a lot of double work, it’s a great way of getting to know new people online!

Pick use cases that are less information-critical

Rather than everybody focussing on the most frequently asked questions and updates, why not look for a use case that’s less dependent on critical information?

People are, understandably, scared. Even with all the re-assurance in the world, we’ve entered an uncertain time, and this can take its toll on people’s stress levels. I’ve seen some examples around of people building chatbots that help people to relieve Corona related stress with meditation and light physical exercise.

Photo by Dominik Scythe on Unsplash

Or create a bot for bringing people together online so that they feel less isolated. Bots on how to work from home, how to get your granny on facetime, educational bots, a meme bot about toilet paper to keep people entertained…The possibilities are endless.

Be prepared to maintain your chatbot…for a long time

I’ve mentioned it already: in case of a rapidly changing, high-impact crisis like Corona, it’s paramount that your chatbot content is up to date. Always. For as long as your bot stays online.

And while building a chatbot is about building from scratch, with all the fun and challenges, maintenance takes stamina, because it’s tedious. Trust me, I’ve maintained aviation manuals for 14 years, and it takes a certain kind of madness to do that (a team of great colleagues helps too).

So before you begin, ask yourself whether you really feel up to maintaining a chatbot for a longer period of time.

So…shouldn’t we build bots then?

I’m not saying you shouldn’t. I guess. But if you do, build them responsibly, with the mindset that fits a crisis like this. And of course, it never hurts to build showcases, examples and experiments. Just make sure that your audience knows the status of your bot.

Stay safe everybody ❤

Photo by Maaike Groenewege