Where do the participants come from? How Nava's panel strategy with Cint and Prolific works

The first question a serious buyer asks me is rarely about the AI. It is about the people. Who am I actually talking to, and how do you know they are real? Anyone who has run panel research has a scar from it: a study that came back fast and clean, and was quietly worthless because half the respondents were not who they claimed to be. So before I talk about reach, I want to talk about trust, because reach without trust is just a bigger pile of bad data.

Two ways to reach the right people

Nava Insights gives you two paths to participants, and you choose based on the study.

The first is to bring your own. Your customers, your users, your employees, your own recruited list. You send them in, the AI moderator interviews them, and you keep full control of who participates. For a lot of customer and employee work, this is the right call.

The second is to tap vetted panels when you need people you do not already have. This is where reach matters, and we are deliberate about which panel does which job.

• Cint brings breadth: 300M+ respondents across 150+ countries. When you need general population, broad consumer segments, or hard-to-find combinations in markets you do not normally operate in, this is the reservoir.
• Prolific brings research-grade quality: 200,000+ vetted participants who know they are taking part in research and are paid fairly for it. When attentiveness and honesty matter more than raw volume, this pool is built for it.
• Norstat brings premium Nordic and European depth for enterprise studies. This one is coming soon, and I will say plainly that it is on the roadmap rather than live today.

Different studies need different pools, and pretending one panel is right for everything is how you end up with the wrong people in the room.

The part nobody likes to talk about: fraud

Panel fraud is real, and it has gotten more sophisticated. Bots, survey farms, people speeding through to collect an incentive, the same person wearing five identities. If you have been burned, you were not careless. You were up against an industry of people whose job is to look like a qualified respondent.

So the honest center of our panel strategy is not reach. It is quality control, layered on purpose.

Every panelist is pre-verified before they ever reach your study. That means identity checks to confirm a real, unique person. Eligibility screening to confirm they match who you asked for. And fraud prevention to catch the duplicate accounts, the inattentive speeders, and the automated traffic that plague open panels. On top of the panel's own checks, you can target with 300+ screening criteria, so "physicians in Germany who switched EHR systems in the last year" is a real filter, not a hopeful guess.

You should never have to wonder whether the person you just learned something from was real. That doubt is the thing we built the verification to remove.

There is one more safeguard, and it is the one I would point a skeptical buyer to first.

You only pay for interviews that hold up

With Nava, you pay for completed, high-quality interviews. Not starts, not attempts, not a block of "responses" you have to clean afterward.

That single design choice changes the incentives. The platform is not motivated to push volume through; it is motivated to deliver interviews that actually count. And voice raises the bar on its own. It is genuinely hard for a bot or a survey farmer to hold a real, adaptive, spoken conversation, answer unscripted follow-ups, and stay coherent across several minutes. The format itself filters out a lot of what slips through a click-through survey. When the deliverable is a real conversation rather than a row of checkboxes, low-effort fraud has far fewer places to hide.

Put together, the model lines up your interests with ours. You want real people and real depth. We only succeed when you get exactly that.

How to choose, in practice

If I were advising you on a first study, I would keep it simple.

1. Already have the audience? Bring your own participants and keep full control of the list.
2. Need broad reach or an unusual market? Use Cint for the 300M+ breadth across 150+ countries.
3. Need maximum research quality at moderate scale? Use Prolific's 200,000+ vetted participants.
4. Enterprise Nordic or European depth? Norstat is coming soon, and we will tell you the day it is live.

Then let the verification, the 300+ screening criteria, and the pay-for-completed model do the quiet work underneath. Reach gets you to the right country and the right segment. Quality control gets you to a real person inside it.

That is the whole strategy, and I would rather state it plainly than dress it up. Big numbers are easy to print. What we are actually selling is the confidence that the voice on the other end belongs to someone who is exactly who you needed, speaking for themselves. Everything else is just plumbing in service of that.