Sociologists from the Pole bureau have integrated AI into their research processes to automate data collection and preliminary analysis. The system utilizes Telegram-based bots that conduct interviews, follow up with respondents, and process responses into structured thematic blocks.
The bureau employs a symbiotic model where AI handles data gathering and initial synthesis, while human experts interpret the results and finalize reports. This approach allows the team to reduce research timelines from two or three weeks to three or four days.
The internal tool relies on Anthropic models combined with traditional algorithms and scripts. Human researchers define the project parameters, context, and guides, ensuring the AI operates within specific constraints to minimize hallucinations and irrelevant outputs.
The team emphasizes that AI is most effective for straightforward tasks, such as measuring product satisfaction or gathering service feedback. They caution that complex topics requiring deep empathy remain challenging for current LLM-based systems to handle accurately.
