What is a synthetic persona? Explaining the synthetic research method built on your own buyers

Learn more about how synthetic personas simulate audience segments you can query on demand between studies—without ever going back to the field.

A synthetic persona is one of four methods filed under the label of synthetic research methods, alongside synthetic boosting, synthetic modeling, and synthetic twins. It also might be the most misunderstood.

If you want to learn more about the full taxonomy, check out our field guide to synthetic data in B2B research. But here, we’ll focus solely on what synthetic personas are and when to use them in your research.

What a synthetic persona is

A synthetic persona simulates an audience segment from research data, queryable on demand without running a new study. It’s a persistent model of a segment you can interrogate over time as opposed to a single survey result frozen in place.

Segment is the operative word here. While a synthetic twin models one specific individual, a persona models a defined audience and forecasts how that audience would respond to a question or direction it hasn’t been asked about directly.

From a single interface, you can run it like a qualitative interview, probing reasoning and language, or like a quantitative survey with structured outputs across a defined question set. The persona extrapolates from real patterns in the data it was built on, and every answer traces back to the research behind it.

What the model draws on

The build starts with the client’s own research: past surveys, transcripts, segmentation, and historical data. That’s the anchor. The model reflects your actual buyers—not a generic stand-in for them. 

NewtonX enriches that foundation with our verified professional data: a network of more than a billion professionals across 140+ industries, with every one identity-verified before entering the panel.

That last point matters significantly in B2B. Consumer persona vendors can train on public data because consumers leave public trails a model can learn from. But B2B buyers don’t post their procurement criteria anywhere public. A model trained on public data produces outputs that sound like B2B reasoning without being grounded in how B2B professionals actually make decisions. The verified professional data layer is what closes that gap.

Not every synthetic persona is built the same way

The difference between the two types decides what you can do with the output:

Marketplace models are pre-trained on shared data any buyer can access. They provide fast, directional reads on audiences that are already well documented, but they can’t reflect your specific buyers, and any competitor can build the same thing.

Proprietary personas are built on your own research and segmentation. They’re locked to your organization and reflect your actual market rather than a generic version of it. This is the model NewtonX builds, and it gets better with use: Each new study you run can retrain the persona and sharpen it, and a sharper persona makes the next study more focused.

When to use synthetic personas, and when not to

The clearest way to think about fit is where you are in the research cycle.

Before your next study: Stress-test survey logic before fielding, check a concept for the objections you’ll need to answer, or sharpen messaging ahead of a campaign.

Between studies: Answer questions that can’t wait six weeks, or explore how a defined segment would likely respond to something that’s shifted since the last wave.

After your last study: Chase the follow-up questions that surfaced during analysis without going back to the field.

But synthetic personas are the wrong tool when the decision has to hold up to outside scrutiny, like a board review or an external audit. The same goes for when a concept is genuinely new with no prior behavior to learn from, or when the question needs lived experience only a real conversation surfaces.

The honest limit

A persona can’t tell you how buyers will react to a product that didn’t exist when it was trained. For decisions that need to be statistically defensible, primary research is still the right call.

What separates a well-built persona from a generic AI tool is how it handles its own limits. It’s built to flag when the underlying data is thin and to point toward the primary research that would fill the gap, rather than returning a confident answer the evidence doesn’t support. The strongest research teams use personas to absorb the everyday questions so primary research is reserved for the decisions that genuinely need it.

Learn more about NewtonX Synthetic Personas here.