Artificial intelligence is substantially transforming how pollsters collect public opinion, with a French emerging company called Naratis leading the charge into what promises to be a faster, cheaper alternative to conventional polling approaches. The company, established in 2025 by 28-year-old engineer Pierre Fontaine, utilises conversational AI agents to perform detailed conversations with respondents, eliminating the time-consuming work that has long characterised qualitative research. Rather than requiring respondents to select options, Naratis’s AI interacts with people in genuine dialogue designed to explore not just what they think, but how they think. The technology purports to provide results significantly quicker and at a fraction of the expense of conventional polling, whilst preserving 90 per cent accuracy—a significant breakthrough as the polling industry grapples with plummeting response rates and growing public distrust.
The Growth of Interactive Polling
At the core of Naratis’s innovation lies a deceptively simple concept: replacing the transactional nature of traditional surveys with genuine conversation. When a participant answers the phone, they meet a youthful, energetic AI voice posing open questions about politics, society, and their personal views. Rather than mechanically recording answers, the system conducts genuine conversation. Three distinct AI agents work simultaneously in the background—one making sure the participant stays on topic, another seeking further understanding when answers seem superficial, and a third verifying the person is genuine and not a bot gaming the system. This multi-layered strategy converts polling from a routine box-ticking task into something far more sophisticated and insightful.
The efficiency improvements are impressive. Traditionally, qualitative research necessitated lengthy and demanding work: recruiting small panels of respondents, conducting individual interviews, converting discussions into text, and then reviewing data for recurring themes and significance. Naratis compresses the timeframe using what Fontaine describes as “parallelisation”—numerous AI tools running interviews in parallel rather than human interviewers working sequentially. A study that once required weeks and tens of thousands of euros can now be finished in one or two days. Responses often arrive within 24 hours, enabling political movements, state institutions and entities to react to unfolding events and evolving public sentiment with minimal delay, substantially altering the speed of polling work.
- AI agents carry out concurrent interviews with several respondents
- Real-time analysis flags shallow answers demanding more thorough examination
- Fraud prevention prevents automated responses and inauthentic answers from skewing data
- Results delivered within hours rather than multiple weeks of conventional research methods
Pace and Effectiveness Transform Survey Research
The survey sector confronts an existential crisis. Response rates have plummeted from over 30% in the 1990s to under 5% today, as noted by AI consultant Stéphane Le Brun. This dramatic decline has created a vicious cycle: lower participation mean increased expenses per completed survey, which in turn renders studies less representative of the broader population. Confidence in polling has eroded in turn, with many viewing surveys as intrusive or unreliable. Against this backdrop, conversational polling powered by AI provides a potential solution, potentially reversing years of falling participation by making the research process itself more engaging and interactive.
Naratis claims its AI-powered methodology achieves results that are “10 times quicker, 10 times more cost-effective and 90% as accurate as human polling.” These numbers, if independently verified, would represent a seismic shift in how organisations understand public sentiment. The financial savings by themselves are transformative: a comprehensive qualitative study that once required tens of thousands of euros and multiple weeks of labour can now be completed for a fraction of the price in days. This democratisation of access could enable smaller organisations, grassroots campaigns and community groups to conduct rigorous opinion research previously available only to well-funded institutions.
Parallelisation: The Key Breakthrough
The innovation driving these gains is refreshingly simple: parallel processing. Rather than human interviewers performing interviews sequentially—one conversation after another—AI agents function concurrently across numerous respondents. This increase in throughput without equivalent expense growth reshapes the economics of polling. Where standard qualitative approaches demanded considerable time and resources, AI-driven approaches reduce timeframes whilst reducing expenditure, permitting businesses to obtain rich, detailed understanding on demand.
Accuracy Claims and Sector Doubt
Naratis’s assertion that its AI methodology delivers 90% accuracy in line with human polling has understandably drawn criticism from recognised experts. The polling industry, built on decades of procedural improvement, remains sceptical of claims that machine learning can reproduce the subtle discernment of skilled researchers. Critics question whether conversational AI can genuinely identify the subtle social cues, hesitations and body language that seasoned analysts use to probe deeper into respondent motivations. The company has not yet published peer-reviewed studies substantiating its accuracy claims, leaving independent verification pending.
Beyond accuracy concerns, industry observers worry about potential biases built into AI systems themselves. If the algorithms powering Naratis’s conversational agents are developed using biased data sets or programmed with unexamined assumptions, those flaws could systematically distort results across thousands of interviews. Additionally, respondents may alter their behaviour when speaking to machines rather than humans, either growing more forthright or more guarded based on their comfort with technology. These psychological and technical variables remain largely unexplored territory, and their effect on polling reliability stays unclear.
- Third-party assessment of precision assertions is awaiting completion from established research institutions
- Potential algorithmic biases could consistently skew results across large-scale AI polling operations
- AI-human engagement dynamics may alter how respondents express genuine opinions and beliefs
The Synthetic Data Challenge
As AI polling expands, a troubling question emerges: how will regulators and the public differentiate between genuine human responses and artificially generated data generated by the very systems conducting the polls? The efficiency and speed that makes AI polling desirable also generates possibilities for manipulation. If an unscrupulous operator were to supplement real responses with computer-generated data, the compiled data could look statistically solid whilst bearing little resemblance to actual what people actually think. The algorithmic obscurity compounds this risk—most voters would struggle to understand how algorithms aggregate and authenticate responses, making it hard for them to rely on the conclusions driving political conversation.
Naratis maintains its systems include fraud prevention systems, with one AI agent designated with determining if respondents are real people or automated systems. However, this safeguard itself relies on AI making determinations about AI, creating a circular vulnerability. As conversational systems become increasingly sophisticated, distinguishing genuine human conversation from synthetic responses may prove technically unfeasible. The opinion research field has long enjoyed widespread credibility partly because its processes are fundamentally transparent—people respond to surveys, findings are compiled. AI polling risks compromising that openness, replacing transparent procedures with algorithmic black boxes that most cannot effectively scrutinise.
Trust and Regulation Concerns
Regulators throughout Europe are only now come to terms with AI’s role in opinion research and political polling. Currently, limited safeguards govern how AI systems collect, process and report polling data. Lacking strong oversight frameworks, the industry faces a loss of public trust if false data enters published results or if computational biases distort findings. French data protection regulators and the EU’s AI Act regulatory bodies must urgently create standards ensuring openness, auditability and oversight in AI-driven polling operations before the technology becomes embedded in political processes.
The Blended Evolution of Opinion Research
Despite the efficiency improvements AI polling offers, industry specialists suggest that human and machine-driven studies will probably coexist rather than one replacing the other entirely. Conventional polling approaches have weathered decades of scrutiny and remain embedded in political institutions, regulatory frameworks and public understanding. Companies such as Naratis recognise that AI performs exceptionally well in speed and cost efficiency, yet human interviewers bring irreplaceable nuance—the ability to read subtle emotional cues, adjust questions instinctively and establish connection that encourages candid responses. A balanced approach combining both methodologies could produce deeper understanding whilst maintaining the transparency voters increasingly demand from research influencing electoral discourse.
The move towards hybrid models, however, demands thoughtful balance. Pollsters must establish clear protocols for how AI-collected information should be balanced alongside conventional methods, and the manner in which results should be communicated to ensure the public understands which methods yielded which conclusions. Training a new generation of researchers to operate proficiently alongside AI systems presents another challenge, as does establishing professional standards that govern the technology’s application. If approached strategically, this evolution could reinvigorate polling practices by enhancing efficiency and reach whilst preserving the human discernment and responsible governance that uphold democratic discourse.