Artificial intelligence (AI) disrupting market research has been an ongoing story in our industry for the last few years. So far, it has mostly been a story about classifying and understanding unstructured data such as text and images. Automating previously human-made processes, reducing costs and improving the delivery speed at the same time. No question that this is a big story. But we think there is a more interesting one.
It’s the story of how our industry can become more human by using AI to better understand human behaviour. A story about how our relationships with consumers, data, and clients will be transformed.
Take consumers for starters
One of our main instruments for data acquisition – the survey – has experienced several changes in the last decade. We’ve migrated from paper to the web. However, although the medium has changed, the content has remained what it was during the last millennium: a set of mostly boring questions asked in a mostly robotic way; a method with not a lot of room for surprises. The almost complete absence of open-ended questions leaves no place for improvisation.
We’ve made humans behave in an algorithmic way when filling a questionnaire. We’ve forced consumers to adapt to our measurement instrument instead of adapting our method of questioning to human behaviour.
AI offers the opposite: the potential to make machines behave more like humans. It allows us to make qualitative research-like studies at scale and to understand the who, what, why, and how in an open-ended question using advanced natural language processing.
AI also enables us to talk to consumers with a more human voice. Creating interviewers that can improvise according to different answers through chatbots, at last having insightful, non-linear conversations with consumers.
Analysts will be able to see data with new eyes
One of the main advantages of AI resides in providing more human representations of knowledge. An image is not a set of pixels, it’s an object or a person performing an action. A comment is not just represented as a set for characters – it has sentiment and intention. A trend is not a set of numbers – it’s a behaviour with a context.
These new representations can be used by the analyst to uncover new insights and understand data in a different way.
How we deliver data to clients will also be affected
Machines are good at understanding tables full of numbers. Humans are not. How can we make more human interfaces to data?
Providing graphic and conversational interfaces for clients to access their data and boost insight discovery is something AI can already do. No more dashboards that only display information: we should build analytical instruments that analysts can play with – instruments that enhance the analyst’s cognitive abilities and allow them to analyse millions of data points easily.
I am not speaking of the future. All these technologies are available right now. Why aren’t we seeing them everywhere? They have been restricted to those with a freakish knack for manipulating abstract symbols. That should end.
Democratising access to these technologies across the company is key. It is when every member of an organisation has quick and easy access to the new technologies that creativity flourishes and transformation lands. Creating the systems and applications that enable this democratisation has to be the main challenge and job of our data scientists.
We have a once in a generation opportunity to disrupt our industry and make it more human. Let’s seize it.
