Ah chatbots! Can't live with them. Can't live without them. And certainly, can't get them to be adopted in a meaningful way by the enterprise. Like every other technology before, enterprises have struggled to get their retail customers and their own internal users to adopt chatbots. Sure, 60% of US adults online use messaging , voice or video chat channels. Voice is an expensive medium - chat less so, but the overall satisfaction with chatbots is still sub-par. Everywhere you look there are articles on how to increase the adoption of chatbots. Everywhere. And they mostly ask you to look for good use cases, look at user-centric design and invest in automation. Here is an easy three step process you can follow to increase your chatbot adoption:
- Get a hammer. A nice ball peen hammer would do
- Break down every little cubicle and office wall till you have an open floor workplace
- Watch your chatbot adoption and your chat channel adoption soar through the roof.
I know the above solution (which I’m calling digital-feng-shui) sounds ridiculous but then the data isn't wrong. Every one of our customers where chat adoption has been better than the average, has invariably had an open working place. So, what is it about the open workplace that drives chatbot adoption? The answer lies in human psychology and culture. Wherever the social cost of having a loud conversation or doing personal work is prohibitive, end users invariably turn towards a chat channel. It's quiet. It's non-intrusive. You can alt-tab to your work screen as soon as your boss decides to walk down the hallway. In fact, if you can segment your traffic by location and by function, do it, and tell me what you see. Invariably we see offices (where there is a high social cost to having loud conversations) veer towards chat. Sales offices, irrespective of where they are, invariably prefer voice. Your seem to influence channel adoption much more than any functionality you can possibly implement. If that is the case, then why does the research invariably tend to suggest otherwise?
The answer to that is cognitive bias. Specifically, the availability heuristic. This is when you provide a lot more weightage to the data that you have, than to the data you don’t. Digital mediums are by their very nature intangible. You can no more feel the bits and bytes of your chatbot than you can feel ghosts. So digital vendors resort to data. And the data only measures the aspects that they can control. Like the rate at which your users end up on a branch of your chatbot, when they give up on the conversation, or when they initiated it. But there are factors that your digital tools cannot begin to measure - like how well designed your office is, or the dynamics of your team. So they don't and all you are left with is slightly accurate but extremely misleading data - and we are all data driven organizations right?
So, what's the solution? How do you track the other aspects of your workplace culture that may deter digital adoption? A simple test that we use is called the hallway usability test. We just point to the first person that we see in the workplace and walk over to their desk. Ask them to use the latest in our digital portfolio. And see how they interact with it. Chances are they don't like it, their environment is too noisy for them to use it, their desks are too small to hold it, or they just don't like the color scheme. But it allows you to see how your latest digital initiative works in the real world. At no extra cost. Hallway usability tests are de rigueur within the consumer ecosystem, where the difference in whether a user likes it or not can mean thousands of dollars. But in the enterprise ecosystem it just translates to two more boring governance meetings and hand wringing around missed KPIs. But it doesn't have to be that way. The next time you need to implement your brand new digital initiative, go do a hallway usability test. If not, just hire that feng-shui expert and get a ball peen hammer!
Reference
[1] https://go.forrester.com/blogs/17-04-20-the_data_digest_chatbots_cant_fully_replace_humans_just_yet
