What changes when your buyer is always available?

What I heard, and what it means, from our session with Bain & Company and Zoom at Cannes LIONS 2026

By Sascha Eder, CEO, NewtonX | June 2026 | Cannes, France

At Cannes this year, I sat down with Kimberly Storin, CMO at Zoom, and Jamie Cleghorn, Global Head of Customer Practice at Bain & Company, for a conversation about synthetic research and what it changes for marketing teams.

The question behind the session was simple: what would your team do differently if you could ask your buyers a question today instead of six weeks from now?

That gap matters more than most teams admit. By the time a traditional study is scoped, fielded, and analyzed, the product may already be live, the campaign may already be out in market, and the decision may already be behind you.

That was the real conversation on stage. Not whether synthetic research is interesting. Whether it is useful. Where it earns its place. And where it absolutely does not.

Synthetic research has a job to do

One of the most common mistakes I see is treating synthetic research like a cheaper version of primary research. But that misses the point.

Kimberly put it well. At Zoom, marketing supports a portfolio with more than 2,500 products and features launched in a single year. No team can run a full primary study for every product question, every message test, every market, every time.

As she said on stage, “synthetic research does not eliminate the need for real customer verbatims or high-stakes research, but it enables teams to move faster and work with greater confidence — even when that confidence is directional. It gives marketing teams an opportunity to become more agile.”

The strongest case for synthetic research is not “run everything this way now.” It is “stop leaving the smaller but still important questions unanswered.” It gives teams a way to pressure-test a message, explore a hypothesis, or get a read on buyer reaction without defaulting to instinct every time.

Jamie made a similar point from the consulting side. Too often, research gets handed over as a static deliverable instead of becoming something teams actively work with across the business.

The bigger shift is organizational

The most important idea from the session was not about efficiency. It was about proximity to the customer.

In many large organizations, customer understanding still sits inside a dedicated insights or research function. Sales may talk to customers. Service sometimes does too. But the people making product and marketing decisions every day often work at a distance from the audience they are supposed to represent.

Jamie’s argument was that synthetic personas can help close that distance by making customer perspective more continuously available inside product development and product marketing.

As Jamie put it, “synthetic personas can bring the customer into the product-development and product-marketing process much more continuously. They can help increase the organization’s collective empathy and make the customer less of something to be studied periodically and more of a perspective that teams can interact with throughout their work.”

There is a real difference between commissioning customer understanding and working with it in real time. One is episodic. The other starts to change how teams think.

Kimberly shared one of my favorite examples from the session. “I found that our marketing teams were very focused on being the voice to the market, but not always as focused on being the voice of the market. We gave marketers several ways to engage directly with customer perspectives — they could walk across the street and speak with a dentist who uses Zoom Phone, conduct social listening, use synthetic research, or speak with an analyst. My expectation is that every marketer will take part.”

A marketing team cannot represent the customer well if it rarely hears from the customer.

Marketing roles will change

Another part of the discussion stuck with me: if customer insight becomes more available, the work around it changes too.

Kimberly talked about the rise of go-to-market engineers and orchestrators inside marketing teams.

“As we see more go-to-market engineers and more people acting as orchestrators — building and coordinating agents — I think a new role will emerge within integrated marketing. That role will involve building, managing, monitoring, and optimizing those agents. We will still need people who understand strategy and the media mix, but we will also need marketers with an engineering-like mindset.”

The best marketers will still need judgment. They will still need taste. They will still need to know what matters to the business. But more of them will also need to understand how to work with systems, not just campaigns.

Synthetic research will be one input among many. The real advantage will go to the teams that know what question to ask, when to trust the output, and when to challenge it.

The risk is losing your strategic center

We also spent time on what can go wrong. Both Kimberly and Jamie came back to the same concern: teams that hand too much over to AI and quietly lose the strategic layer above the work.

“One risk is that we delegate everything and reduce the work to a prompting exercise without maintaining a North Star. If you are not asking the right questions, critically evaluating whether an answer makes sense, or determining whether you are testing the right hypothesis, you end up with people who are purely execution-oriented rather than strategic.”

Bad tools are a problem. Overconfident teams are a bigger one.

Jamie’s answer was more cultural, and worth keeping in full. “The real question is how you make customer centricity part of the organization’s everyday behavior rather than a separate exercise. Now that teams can interact with a customer perspective more continuously, that perspective can be inserted directly into regular workflows and conversations. You could imagine beginning a staff meeting by asking what the customer would say or using that perspective to challenge the assumptions being made.”

That shift matters more than any individual tool.

Build or buy is the wrong first question

I asked both Kimberly and Jamie how enterprise teams are evaluating synthetic research solutions today.

Kimberly described Zoom’s approach as AI-first. “At Zoom, our first question is always: can we build this ourselves? Zoom is an AI-first platform, and generative AI is embedded across our workflows. We began by testing what we could create ourselves. We have built strong internal agents for content and product marketing, and they have meaningfully accelerated our process. Right now, we are still testing whether what we build internally will be good enough. At the same time, we have begun speaking with vendors and exploring external options. Ultimately, we would work with a partner that can do it better, faster, or more cost-effectively.”

Jamie had a similar answer. “The model will likely become more continuous and hybrid. Instead of relying only on occasional studies, companies will be able to build a longer-term relationship with customer understanding and return to the same questions and audiences over time. The stronger model will combine synthetic interaction with real-customer research, using each where it is most valuable.”

The build-or-buy conversation gets a lot of attention. The harder question is what data sits underneath the system.

Synthetic outputs inherit the strengths and weaknesses of their training data. And in B2B, that matters even more. Generic public-data models do not naturally understand B2B buyers because those buyers do not leave behind the same volume or type of public signals that consumer models feed on.

That is why data quality remains the center of the conversation.

The real differentiation question is still ahead of us

I ended by asking where this goes over the next three to five years.

Kimberly answered that she was most worried about commoditization. “Whether companies are using OpenAI or another platform, everyone increasingly has access to the same underlying capabilities. The question is: how do we build something that is unique and proprietary to us — something different from what our competitors are doing and that gives us a real advantage? Right now, the technology is still new enough that any gain in efficiency or agility feels like progress. But three to five years from now, when the playing field has leveled, the real question will be whether we have built anything meaningfully differentiated.”

I have been thinking about this, too, at NewtonX.

Anyone can build a model with AI. The harder part is feeding it something real. Something specific to your buyers. Something your competitors cannot replicate with a prompt and a public dataset.

That is where the next advantage will come from.

What I took away from Cannes: the data advantage

I left the session with a few convictions.

Synthetic research is valuable when it helps teams stay close to the customer between larger studies. It keeps decisions moving.

The biggest upside has little to do with cost. It comes from making customer understanding more available across the organization.

The danger is not the tool itself. The danger is using the tool in place of judgment.

And over the next few years, the real differentiator will be the data underneath the system. Not the model alone.

That is the part of this market I would pay closest attention to.

At NewtonX, that has been our view from the beginning. If synthetic personas are going to shape real decisions, they need to be grounded in real, current, buyer intelligence.

About Synthetic Personas 

NewtonX Synthetic Personas are built on identity-verified professional data from across 140+ industries, trained on your own research, and designed to get sharper with every study you run. Learn more about how Synthetic Personas work.