---
id: "framework-ai-moderation-use-cases"
type: "framework"
source_timestamps: ["§ When You Need the “Why” Behind the Numbers", "§ When You Need to See What People Can’t Say", "§ When the Topic Is Too Sensitive for a Human Interviewer", "§ When Respondents Are Hard to Reach or Schedule"]
tags: ["decision-framework", "use-cases", "strategy"]
related: ["concept-llm-based-interviewers", "action-setup-poc", "concept-frontier-listening", "concept-multi-modal-video-insights", "concept-asynchronous-qualitative-research", "claim-ai-reduces-impression-management"]
steps: ["\\\"When You Need the 'Why' Behind the Numbers: use AI to blend qualitative depth with quantitative breadth", "scaling open-ended probing to understand the reasons behind shifting metrics (e.g.", "Microsoft's Frontier Listening).\\\"", "\\\"When You Need to See What People Can't Say: deploy multi-modal", "video-based AI to observe actual behaviors and attitudes in natural contexts (e.g.", "Unilever's kitchen ethnography).\\\"", "\\\"When the Topic Is Too Sensitive for a Human Interviewer: use AI to reduce social friction", "fear of judgment", "and impression management on sensitive topics like health conditions or personal insecurities.\\\"", "\\\"When Respondents Are Hard to Reach or Schedule: implement asynchronous AI interviews to capture insights from time-poor", "high-value audiences (doctors", "executives) who cannot commit to live scheduling.\\\""]
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"
---
# Framework for Assessing AI-Moderated Qualitative Interviews

The authors provide a practical, four-part framework for business leaders to evaluate where AI-moderated qualitative interviews add **immediate** value. It maps the optimal deployment of [[concept-llm-based-interviewers]] onto four situations where traditional methods fall short. Before scaling, pair it with [[action-setup-poc]].

## The four use cases

1. **When you need the "why" behind the numbers** — blend qualitative depth with quantitative breadth; scale open-ended probing to explain shifting metrics. → See [[concept-frontier-listening]] ([[entity-microsoft-d5]]) and the efficiency case in [[claim-sweetgreen-efficiency-gains]] ([[entity-sweetgreen]]).

2. **When you need to see what people can't say** — deploy multi-modal, video-based AI to observe behaviors and attitudes in natural contexts. → See [[concept-multi-modal-video-insights]] ([[entity-unilever-d5]] × [[entity-conveo]]) and [[claim-ai-captures-unspoken-behaviors]].

3. **When the topic is too sensitive for a human interviewer** — use AI to reduce social friction and impression management. → See [[claim-ai-reduces-impression-management]] and [[contrarian-ai-better-for-sensitive-topics]] ([[entity-chubbies]], men's-health/[[entity-outset]]).

4. **When respondents are hard to reach or schedule** — use asynchronous AI interviews for time-poor, high-value audiences. → See [[concept-asynchronous-qualitative-research]] ([[entity-doximity]] × [[entity-outset]]) and [[claim-ai-reaches-unavailable-audiences]].

## How to use it

Treat the four cases as a **diagnostic checklist**: identify which failure mode of traditional research you face (missing "why," unobservable behavior, sensitivity friction, or scheduling impossibility), then select the matching AI-moderation mode. Enrichment adds a fifth discipline the article underplays — invest in **research design and oversight** (see [[action-establish-metrics]]), because scaling a poorly designed instrument merely multiplies biased questions across thousands of sessions.