User Research
User Research Insights: Survey Results on Human vs. AI Personal Assistants

We recently did some user research in planning and managing people’s schedules for is a virtual travel assistant, which automatically adds travel time and details to user’s digital calendar.

One of the interesting insights we got from this research was finding out what people were comfortable with an Artificial Intelligent Personal Assistant (AIPA) to control and what people would prefer a human assistant to be responsible for as well as a blend of both.

Survey Results

We conducted qualitative and quantitative research to explore user views toward planning and scheduling and the use of AI assistants. Here are the results from the survey element of the research.

The survey respondents spanned across 44 people, with 55% female and 45% male, with the majority aged between 26–35 (61%) and 20% aged between 36–45 years.

We found that when it came to using human personal assistants, 18 respondents had used a personal assistant. Of these 18, 56% had worked with an employee of a company and 17% had worked with a virtual personal assistant.

For those 26 that hadn’t used a personal assistant, it was mainly due to not being able to afford one (46%), not having considered one for home/domestic use (38%) or didn’t feel the need for one (27%).

Here’s where the results get interesting. We asked about personal assistants for daily life and what people would like humans, AI assistants and a blend of the two, to do.

For daily life, human personal assistants were seen to be potentially beneficial for calendar management (64%), travel management (57%), scheduling meetings (55%) and managing to-do lists (48%).

Only 32% of the respondents had heard of an Artificial Intelligent Personal Assistant (AIPA) with 89% of all respondents supporting the idea of using one. The respondents were happy for the AIPA to help with calendar management (68%), manage expenses (59%), schedule meetings (59%) and manage to-do lists (57%). This shows that respondents were happy for the AIPA to help with tasks similar to the human personal assistant, with addition of the AIPA helping with expenses management.

When asked if respondents were more or less likely to use an AIPA, as it was supported by a human, 39% said they would use a human-assisted AIPA, 30% were indifferent, 14% would use a human-assisted AIPA over anything else, 9% would prefer the AIPA and 7% would prefer a human PA.

A Blend of Human and AI Personal Assistants

The survey shows that the respondents were supportive of the idea of using an AIPA. There were similarities between the human and AI personal assistant doing similar tasks, predominantly calendar management, scheduling meetings and managing to-do lists. This can be seen by products such as and that primarily do these tasks.

Although there was support among the majority of respondents, around 40% would prefer to use a blend of an AI and human personal assistant compared to a human only or an AI assistant only.

We found in the qualitative work that users liked to have some autonomy over what the AI assistant does. Having a blend of AI and human personal assistants could mean that the individual could get more value from the personal assistant if some of their tasks were automated.

This finding resonates with some work Snap Out has previously has done on chatbots with Virgin Trains, where some of the messages were automated but the social media team sense-checked these before they were sent out.

If you’d like to work with us to learn more about what your customers what from your AI, drop us a line.

The Snap Out Team ?