---
id: "concept-llm-based-interviewers"
type: "concept"
source_timestamps: ["¶ 3", "§ When You Need the “Why” Behind the Numbers"]
tags: ["generative-ai", "research-methodology", "ai-moderators"]
related: ["concept-scaled-empathy", "concept-asynchronous-qualitative-research", "claim-ai-resolves-research-tradeoff", "entity-anthropic", "framework-ai-moderation-use-cases"]
definition: "AI agents powered by large language models that conduct dynamic, in-depth qualitative interviews with human respondents at quantitative scale."
sources: ["commercial"]
sourceVaultSlug: "hbr-seg-commercial"
originDay: 5
articleStem: "hbr-new-30-ai-scale-customer-research"
sourceUrl: "https://hbr.org/2026/04/how-ai-helps-scale-qualitative-customer-research"
sourceTitle: "How AI Helps Scale Qualitative Customer Research"
---
# LLM-Based Interviewers

**LLM-based interviewers** are AI agents designed to conduct in-depth qualitative research conversations with human respondents. They overcome the traditional breadth-versus-depth tradeoff in market research by engaging large numbers of respondents (comparable to quantitative surveys) while capturing the nuance, context, and interpretive richness of human-led interviews. Mechanically, they combine **predetermined open-ended questions** with **dynamically produced follow-ups** generated in real time by the LLM.

They are already being deployed at global scale: [[entity-anthropic-d5]] used a Claude-based "Anthropic Interviewer" to conduct **over 80,000 interviews in 159 countries and 70 languages** (see [[quote-anthropic-scale]]). This concept is the substrate for [[concept-scaled-empathy]], underpins [[claim-ai-resolves-research-tradeoff]], and enables downstream modes like [[concept-asynchronous-qualitative-research]]. The four canonical deployment situations are catalogued in [[framework-ai-moderation-use-cases]].

## Calibration for a downstream agent

Enrichment sources corroborate the core capability — practitioners describe this as "qual at scale" and note dynamic follow-ups that probe the "why." However, the specific Anthropic figures (80k / 159 / 70) are **company-reported and not independently audited**; treat them as order-of-magnitude illustrations of what concurrent AI conversations make possible rather than verified benchmarks.


## Related across articles
- [[entity-agentic-ai-d5]]
- [[concept-digital-modalities]]