What users want when they use a voice bot.

Here are some questions to consider when dealing with developing a voice bot.
First, without much surprise, the audience prefers to deal with a voice that does not seem synthesized, but that is fully human (70.8%)
It should be noted, however, that in the 30 to 59-year-old group, those who prefer a human voice drops to 66.1%. Which probably means: I prefer to know who I’m talking to.
On the other hand, there are no big differences if you analyze the preferences regarding the gender of the voice: in the case of the natural human voice, there is only a 2.3% difference between users who prefer the natural female voice and those who prefer the male voice (against the latter).
The percentage difference grows to 12.5% in favor of the female synthetic voice over the male synthetic voice, indicating that the female synthetic voice is more tolerated than the male voice (note that both male and female users prefer the synthetic female voice).
Another important factor for a good UX is understanding how long the voice bot dialogue must be before it gets boring for users.
A vocal interaction divided into two parts was used in the analysis: an introduction and a text/list.
For the groups were subjected 4 UX in which they were alternated: short introduction and list, long introduction and list, long introduction and short list, and short introduction and long list.

These are the results:

  • 81.8 percent found the first experience (long intro and list) too long;
  • in the case of a short introduction and long list, 50% found the UX adequate, 50% too long;
  • in the case of long introduction and short list, 55.6% found the UX adequate, 44.4% too long;
  • if there were an intro and short list, 54.5% found the UX adequate, 27.3% too long, while 18.2% found it too short.


Here are some considerations that can be made:

  • 18% of users consider the UX with short intro / short list “too short”, which does not happen in other cases;
  • in groups two, three and four (that is, in cases where at least one of the two parts is short) those that consider the adequate experience are between 50-56%, that is, there are no large variations;
  • in groups that both had a short list configuration, the difference between the groups that consider the experience adequate is minimal, demonstrating that the long introduction was more tolerable than the long list, which becomes boring in spoken dialogue.
  • but above all: in designing the conversational architecture of your voice bot, be concise!

#Voicebot #VocalMultichannel #chatbot