Synthetic personas became specialized media coverage: what changes for teams that decide before launching

Meio & Mensagem published a report explaining how synthetic personas simulate consumers’ rational and emotional reactions to test campaigns, products and pricing before launch. For anyone deciding the next launch or campaign, the message is direct: synthetic personas are no longer an experimental bet; they are now part of real decision-making processes.

What does Meio & Mensagem’s coverage confirm about synthetic personas?

The report describes synthetic personas as virtual profiles built from large volumes of real data — mostly zero- and first-party data, refined through quantitative and qualitative analysis layers. AI algorithms identify decision and behavior patterns in that data and generate profiles that can simulate communication, product testing, positioning or a full purchase journey. To go deeper into the concept, read our complete guide to synthetic personas.

Two technical points cited in the article deserve attention from anyone evaluating providers of this type of technology:

• The report states that, in Galaxies’ models, a persona is only considered valid if it reaches 80% accuracy compared with real consumer responses. This means the persona is not generic: it has a specific objective and temporal validity, calibrated to the project context.

• Unlike generic chatbots trained on open internet data, the model is fed with project-specific information — reducing the risk of answers that are disconnected from the target audience’s reality.

HSR’s Head of Innovation, Lucas Pestalozzi, summarizes the practical value of the technology as a way to reduce information gaps and accelerate decisions that traditionally took weeks to reach a reliable answer.

Do synthetic personas really work? The numbers Galaxies has already validated

The report mentions companies that used synthetic personas to validate campaigns, evaluate product experiences and simulate consumer journeys, without detailing numbers. Two cases we have already published with open methodology are:

• Bradesco Seguros: cost per respondent fell from R$ 17.00 to R$ 1.20 — a 93% reduction in cost per interview, with complete campaign validation in 48 hours, compared with the weeks this type of decision used to take.

• Meta: the goal was to reduce customer acquisition cost (CAC) and accelerate creative optimization. With synthetic personas accelerating part of the decision process, Meta’s team recorded a 52% reduction in CAC and 70% more speed in improving creatives.

These numbers do not appear in the Meio & Mensagem article — and that is precisely the value of going beyond press coverage: it confirms the category is real, but whoever decides budget needs auditable methodology, not just editorial approval. Explore more client cases and content about market research ROI.

What the report does not detail: how to evaluate a synthetic persona provider

The article mentions, in general terms, that using the technology requires attention to ethics and data governance, with practices such as deep eval validation and guardrails to reduce bias. That is true, but insufficient as a buying criterion: any provider can claim to follow best practices. What changes the conversation is asking to see evidence.

In practice, we recommend that marketing teams require any synthetic persona provider — Galaxies included — to answer four objective questions before approving a project:

• It is not enough to hear “our persona is accurate”. You need to know which real response base the model was validated against and what minimum percentage was accepted.

• A synthetic persona built on generic internet data tends to answer generically. A persona built on zero- and first-party data from your audience answers specifically to your context.

• Encryption and access control mentioned in press coverage are declarations; a certification such as ISO 27001 is third-party verification.

• A client quote is social proof. A case with respondent count, unit cost and delivery time is methodological proof.

That last point is what separates research content from advertising content disguised as research — and it is why we publish the numbers from the Bradesco Seguros and Meta cases with methodology, not only results. To understand the security pillar, read about Galaxies’ ISO 27001 certification.

Why is now the right time to adopt synthetic personas in Brazil?

Entering a new technology early often sounds risky — but waiting can cost more than acting. The Meio & Mensagem report itself confirms that marketing and communications, product development, retail, financial services and technology are already among the first sectors benefiting in Brazil, with a trend toward integrating generative AI and hybrid market validation methodologies.

The technology is already validated for real marketing decisions: brands such as Nestlé, Magalu, O Boticário, Bradesco Seguros and Meta use it in production, not as a pilot. What is still on the growth curve in Brazil is the number of marketing teams that already know how to interpret this type of data with confidence — and that is exactly the advantage of deciding now.

For the CMO, the practical point is direct: whoever starts reading this type of signal today reaches the next launch with a sharper view than the competitor who only reacts to the trend when it is already obvious to everyone.

How can you apply this to validating your next launch?

If you are planning to launch a product, reposition a brand or test a campaign in the coming months, the practical question is not whether synthetic personas are legitimate — specialized media coverage has already answered that. The question is: at what stage of the decision do they enter, and what do they prevent you from discovering too late, after media has already been bought?

In our experience with brands such as those cited in the report, the pattern that works is:

• Concept and message testing: synthetic personas enter before any test with real audiences, eliminating weak hypotheses and taking the 2-3 most promising variations into final validation.

• Pricing and positioning: simulations make it possible to test multiple price ranges and naming options in parallel, without the incremental cost of recruiting new groups for each variation.

• Journey and cognitive barriers: the conversational format of synthetic personas reveals where the message causes confusion or objection before the campaign goes live — not after, when correcting it has already consumed media budget.

The point of attention raised by the report, and reinforced internally by us, is governance: encryption, access control and compliance with LGPD and GDPR are not a compliance checklist — they are prerequisites for a synthetic persona to represent a real audience without exposing anyone’s sensitive data. Explore Galaxies solutions to apply this type of validation to your next project.

Frequently asked questions

At what stage of a product launch do synthetic personas help the most?

In concept, message and pricing tests, before the campaign goes to market. They help eliminate weak hypotheses and identify cognitive barriers and audience objections with the 2-3 most promising variations, reducing the risk of discovering a message problem only after media has already been bought.

What is the minimum criterion for trusting a synthetic persona?

According to Meio & Mensagem’s report, Galaxies’ model only considers a persona valid above 80% accuracy compared with real consumer responses, with temporal validity defined for the project context.

How do you know whether a synthetic persona provider is reliable?

Ask for evidence, not declarations: what the accuracy index is and how it was measured, whether the data is exclusive to your project, whether there is third-party information security certification such as ISO 27001, and whether the provider shows open methodology — respondent count, unit cost and delivery time — not only a client testimonial.

How much does it cost to validate a campaign with synthetic personas?

The cost varies by scope, but in the Bradesco Seguros case, cost per respondent fell from R$ 17.00 to R$ 1.20 — a 93% saving per collected response, with delivery in 48 hours.

Source cited: Larissa Santiago, “Personas sintéticas simulam razão e emoção do consumidor”, Meio & Mensagem, October 8, 2025.