July 2026. Until recently, knowing what the consumer wants was expensive, slow, and perishable. Mid-scale qualitative research used to imply weeks of work, dedicated teams, and budgets that few firms could sustain regularly. That model has just broken.
Artificial intelligence today allows reducing market research costs by up to 90% and compressing timelines that used to be measured in weeks into hours. Analysts project that synthetic data will account for more than 50% of market research inputs in 2027 (PyMC Labs), and 69% of researchers already use this technology, with 83% planning to increase their investment in AI applied to research during 2026 (The Alchemic).
For marketing, innovation, and insights departments at large companies, the impact is direct: faster decisions, more efficient budgets, and a higher level of analysis in the processes for launching new products.
The problem cost-cutting doesn’t solve: research that goes stale
But cutting costs isn’t the only argument. There is a quieter problem affecting any company that has invested in qualitative research: information dies. The report is archived, nuances fade, and, over time, an expensive investigation stops delivering value.
One of the leading consumer goods multinationals operating in Spain experienced it firsthand. It had launched a large-scale study — 21 focus groups, 12 in-depth interviews, and 2 online communities — with nuances critical for strategic launch decisions. The problem was the usual one: synthesizing it without losing richness was unfeasible.
The solution came from Kraz, a Spanish company applying AI to business. To be precise, there is a nuance, because it marks the difference with the trend of synthetic consumers: no virtual consumer was invented here. Its Atria platform trained an AI assistant on the complete transcripts of the real sessions, the sessions of real people. The result is a consultant available at any hour that marketing and innovation teams can ask in natural language and who responds with evidence from real consumers, instantly comparing perceptions across segments.
What previously required weeks of manual analysis is now immediate queries. Decisions about launches planned for 2026 are being directly supported by that assistant.
“A traditional focus group dies when the recorder is turned off: what remains is a report that summarizes and therefore loses nuance,” explains Joan Miró, CEO of Kraz. “What AI does is the opposite of replacing human opinion: it makes it a living asset, consultable forever and without losing richness. The debate isn’t humans or machines. It’s how much value we are letting die in the research we’ve already paid for.”
Kraz’s stance reflects the model taking hold in the most advanced companies: not the replacement of the real consumer with a synthetic one, but the building of synthetic analysts from the opinions of real users. AI on real human data for speed and accessibility, with real people as the source of truth.
A divided industry in the face of a change that is already irreversible
Not everyone in the sector is celebrating the transformation. According to a Rival Group study, 42.75% of researchers say they are “not at all excited” about synthetic respondents, especially due to their ability to capture emotional nuance and cultural context. Validation studies place the correlation between synthetic and human responses at 80% to 95% for directional questions, a margin some in the industry deem unacceptable for strategic decisions.
Yet adoption numbers suggest the debate is no longer whether change will happen, but at what speed and under which model. For companies that push launches with tight budgets and deadlines, the question isn’t technological anymore. It’s competitive.
