Three mistakes large organisations make when implementing a chatbot or voice assistant
Author: Hans van Dam
Co-founder at Robocopy.io
Published on 22 August 2019
We see a clear pattern in how large corporations design their dialogues for chatbots and voice assistants. You can’t really blame them, but they do all make the same 3 mistakes over and over again resulting in a poor functioning chatbot. In this article we give you a little advice to help prevent these mistakes.
Every organization has knowledge managers. Often, these are people who’ve been working there for years and know the company inside out. They are excellent at what they do, however, they to tend to focus too much on the exceptions. For every question they receive the answer needs to be thorough and complete. Even if answer A is helpful to 80% of people, this won’t be enough for a knowledge manager. He or she will invest a lot of time to create answer B in order to help another 10% of people. The same goes for the 5% answer C will help. This goes on and on until even the smallest detail is covered.
Recognizing and acknowledging the exceptions is valuable, but it is not the best starting point for conversation design. Using the Pareto-principle, also know as the 80/20-rule, works better. According to this principle 80% if user questions deal with 20% of the problems. If you aim for this 20%, you will serve 80% of your customers a helpful answer.
A technology driven approach can lead to poor conversations as well. When starting a chatbot or voice assistant, many organizations start with a technical framework. As soon as they have invested in a bot-platform they start designing dialogue within the boundaries of that platform. Because of this approach, developers often think that every problem needs a (time consuming) technical solution. The effect: a design process where tech is leading and creativity is inferior.
A third mistake many organizations make is letting business processes determine their conversation design. Most businesses have a protocol for every process. Of course they need these protocols to a certain extent, but if you simply convert these into flowchart the result will never be natural conversation. Don’t focus on the process, focus on the user.
User centric design means you place natural conversations at the centre of your design process. A great starting point is ‘role playing’. One person plays the user and the other plays the chatbot. Record the conversation and analyse your findings. This forms the basis for the design of a natural conversation. Of course this first dialogue is only the beginning. By applying different copywriting techniques from both psychology and technology, you can create strong dialogue. Useful, natural and persuasive dialogues for your user.
