Analyzing data
Crunching and summarizing large sets of data is one of the first (and clearest) uses of generative AI, and one of the most common. One in four AI users have used the tools for summarizing complex subjects (23%) or organizing data (22%), according to the March 2024 Ipsos poll.
This becomes particularly useful within market research. Previously, if brands wanted to get the opinions of many people, you had two main choices: Ask a poll question with a defined set of answers from which people must choose, or ask an open-ended question where people can write what they like, and spend a lot of time and money for a human to analyze the data in a meaningful matter.
Polling allows you to sort people into buckets more quickly, while open-ended questions allow you to get more complex (but more divergent) answers. However, grouping together the answers to open-ended questions required a human to go through hundreds – perhaps thousands – of answers to identify themes, or otherwise just resulted in a simple output like a word cloud.
But a robot can listen to more people at once. Ipsos has been using machine learning for years to find themes in our research, so for us, this is an evolution, not a revolution. But the new capabilities of generative AI represent a leap forward: It can near-instantly sort through hundreds or thousands of responses to find themes in written text. It can then give you an output that’s far more detailed or nuanced than any word cloud – bringing the voice of the people into your business, in their own words.
