Abstract collage illustration of a human hand interacting with data elements

Human-Centered AI Connects Data with Emotion for Strong Activation

by Shari Aaron Executive Vice President, Growth & Innovation

and Shayna Beckwith Vice President

With two faces, the Roman god Janus represents duality. One face looks to the past, and the other to the future. The key in his hand signifies his authority over transitions, entrances and exits, which is why you’ll often see him cast into archways and over doorways.

This ancient idea is prescient now, as artificial intelligence (AI) evolves, promising human-like intelligence. Janus would most likely recommend that we usher in the new while honoring the traditions of the past. We need a gatekeeper like Janus to help with checks and balances as we transition to an increasingly AI-influenced environment.

Bridging the gap: Building trust into the data.

A machine is still a machine; you have to guide it. There is a strong desire to take advantage of the deep data, speed, and efficiency that AI research promises, but the technology is far from the autopilot stage. Approach each project carefully to ensure AI-based research is accurate and actionable.

There are ethical implications to consider when using AI, as well as the need to mitigate biases in the system’s outputs. Our team carefully monitors and evaluates the findings to ensure data quality. Building trust through transparency and iterating based on feedback further enhances the reliability and relevance of AI outputs.

Quantitative work with AI requires the same attention to structure and detail from data scientists that they’ve always applied when working with data technologies. Human involvement is required to ensure the validity of the research. We begin by conducting stakeholder interviews to understand business goals and define clear objectives. We use data from past research and leverage AI tools to gain new insights, maintaining data quality throughout the project with data checks and frequent validation. Our team interprets AI-generated insights using their experience and knowledge.

A machine is still a machine; you have to guide it. There is a strong desire to take advantage of the deep data, speed, and efficiency that AI research promises, but the technology is far from the autopilot stage. Approach each project carefully to ensure AI-based research is accurate and actionable.”

 

Machines, however, are not able to capture the subtleties of emotion, sensory experiences, and language that are important to understanding what consumers want and why they want it. AI can help us understand which questions to ask in the quantitative research phase. AI tools help us conduct background research and generate brainstorming ideas for interactive sessions with clients. These tools are enriching our qualitative research and help us frame stronger questions for IDIs and look for the right signals in our observational and collaborative work. Input from consumers augments and validates what our broad spectrum of research—from surveys to advanced AI—has taught us.

Human-centered AI research bridges the duality between data and insight.

Our recent work with AI proves something we’ve always known about research: Quality data is central to the success of the project, but consumers will always add the details that separate the wheat from the chaff to turn good research into invaluable insights. Here are a few examples:

  • Ideating product names: Generative AI is ideal for creating lists of potential names based on a few logical inputs, but when we showed consumers the list of the strongest ideas generated by AI in a Real Time Concept Optimization session, participants vetted and updated the product name ideas until they came up with a unique name that resonated with the goals of the project and what they thought would be most useful.
  • Understanding consumer habits across markets: For a confectioner looking to grow their category in global markets, we mined 85 million data points from video, social media posts, product reviews, and more to gain a deeper understanding of each market, and then re-mined the data to quantify findings and support additional insights.
  • Exploring taste and preferences: During an Innovation Sprint,™ generative AI helped us uncover emerging trends in flavors that appeal to Gen Z consumers, informing early-stage idea creation.


Remembering that the “A” in AI stands for “artificial.”

AI is helpful up to a point. It’s ideal for numbers and data crunching, but it gets confused when it comes to creative tasks. For example, we were recently preparing for a brainstorming session with clients and used generative AI to create images and taglines that we could use for mood boards in a collaborative session. Our input for images was “illustrate a family on a couch enjoying candy in different cultures globally,” and while a few of the images were fine, several interpreted “cultures” with simplistic and sometimes insulting tropes. Careful oversight is necessary.

We will never be able to conduct meaningful research for brands based solely on AI research. Humans play the essential role in revealing the subtleties that add texture, color, and meaning that lead to “aha” moments for our clients.

 

Take a dual approach to your research program with our advanced tech capabilities + deep experience.

Contact Us