How leading brands are transforming innovation with AI and connected data

Kelsey Sullivan

This month, Zappi held the annual Breakfast Club event in London with leaders from the insights industry, including Johanna Thompson, Human Intelligence - Agile Innovation Insights Senior Global Manager at Mars and our very own Steve Phillips, Founder and Chief Innovation Officer, Babita Earle, Managing Director, International and McKenzie Davis, Director, Professional Services at Zappi. 

During the event, in addition to discussing our Connected Insights Framework and holding a workshop on the current state of connected insights, these industry leaders shared their learnings and experience with how AI can enhance the innovation process when leveraged with connected data.

If you missed it or are looking to share details with your colleagues, here’s some of the key takeaways.

“We want to get to more of this fully connected world. If we can get there, there are incredibly exciting things we can start analyzing — giving us a holistic picture of the market and the consumer.”

- Steve Phillips, Founder and Chief Innovation Officer, Zappi

1. AI enhances innovation, but human bookends are essential

Generative AI can help speed up product innovation, but it’s not a magic fix. 

In early tests with Mars, AI-generated concepts looked good on the surface but didn’t always connect with consumers, lacking brand recognition and believability. 

The real value came when humans stepped in to refine AI outputs — boosting key metrics like purchase intent and uniqueness, as well as adding their expertise. 

“Generative AI brings potential to product innovation, but needs ample human intervention to thrive.”

- Johanna Thompson, Human Intelligence - Agile Innovation Insights Senior Global Manager, Mars

The takeaway? AI works best as a partner, not a replacement, for human creativity.

2. Experimentation and customization are key to effective AI use

Mars and Zappi tested multiple AI-human collaboration models to understand where AI adds value. They didn’t just assume AI would work — they put it to the test.

“As any good researcher, we started with a hypothesis…and tested multiple scenarios to see how AI performed versus human teams.”

- Johanna Thompson, Human Intelligence - Agile Innovation Insights Senior Global Manager, Mars

They found that structured experimentation, where AI was used for specific tasks yielded the best results, instead of simply dumping everything into AI and hoping for the best. 

“On the earlier experiments, we weren't scoring high on believability or brand linkage because the AI was simply getting overwhelmed with all the documents we fed it. Whereas now we had the Agents with a sole task to to focus on, like brand expression, and this was really great for us to get our hands on, experiment with and see.”

- Johanna Thompson, Human Intelligence - Agile Innovation Insights Senior Global Manager, Mars

Customizing the AI process improved effectiveness. Rather than overwhelming AI with too much data, they assigned AI Agents with specific tasks to produce more relevant and actionable outputs.

3. Proprietary data is your competitive edge

Since most companies are using the same AI models, the real advantage comes from feeding AI with proprietary insights. 

“We all raised our hands and said we had access to LLMs — but basing AI outputs on competitive insights is what makes them truly valuable.”

- McKenzie Davis, Director, Professional Services at Zappi

If you’re only using generic data, your outputs won’t be any better than the competition’s. But by leveraging exclusive consumer insights, brands can guide AI toward ideas that are more relevant, differentiated and strategically valuable.

“Own your edge. Make sure you’re leveraging your proprietary data set. Everyone has access to AI, but your unique data is what gives you a real competitive advantage.”

- McKenzie Davis, Director, Professional Services at Zappi

4. AI is fast, but human judgement makes it smart

AI can generate ideas in seconds, but speed doesn’t always mean quality. 

At first glance, AI-generated concepts can seem polished, but they often lack the depth, nuance or brand connection needed to make an impact. 

“AI outputs look very skillful to the untrained eye. When you have an expert at the table, they will tell you it's scratching the surface, and they can then really push AI to dig deeper and bring something to the surface that's more unique and different.”

- Johanna Thompson, Human Intelligence - Agile Innovation Insights Senior Global Manager, Mars

The real magic happens when AI and human creativity work together — AI speeds up ideation, while experts refine and shape the best ideas.

“AI can generate ideas quickly, but it’s the human nudge — iterating relentlessly, correcting biases and refining outputs — that makes a concept truly strong.”

- McKenzie Davis, Director, Professional Services at Zappi

Wrapping up

When used strategically, AI can significantly enhance the innovation process. However, to truly unlock its potential, AI must be used alongside human expertise. As we move forward, combining the strengths of AI, connected data and human insight will continue to be the key to successful innovation.

“Where we want to get to is much more of this fully connected world. A world where all of your data can and should talk to each other.”

- Steve Phillips, Founder and Chief Innovation Officer, Zappi

We want to thank those of you who attended this event and hope it provided valuable insight into the future of consumer insights. 

For those of you who did not attend, we hope these highlights provided tangible insights into how to best work with AI and look forward to seeing you at the next Breakfast Club event! 

Zappi AI Agents ✨

Learn how you can use generative AI to help you come up with unique ideas at speed with the combined power of consumer data and expert human insights.

If you’d like to learn more about Zappi events or solutions, talk to us.