From Passive Mining to Proactive Matching: A Survey of AI-Driven User Need Discovery and the Case for Agent-Mediated Demand-Supply Coordination

Firms have long struggled to learn what users want before deciding what to produce. This survey reviews research published from 2024 to 2026 on using artificial intelligence (AI), in particular large language models (LLMs) and LLM-based agents, to discover user needs from user-generated content such as product reviews and social media posts. We searched Google Scholar with a bilingual English and Chinese query matrix and applied deduplication, bibliographic verification, rule-based relevance scoring and citation snowballing, which produced a verified corpus of 98 papers. Each paper was classified by technique, application domain and year. Within this corpus, LLMs are the most common technique, agent-based work is growing from a small base, and retrieval-augmented generation is rarely used. Applications are concentrated in software requirements engineering and e-commerce, while other industries remain thinly covered. We propose that this pattern reflects how much public user-written text each industry produces, and we present this as a proposition for future testing because our search design favors such industries. Reported accuracy is promising on narrow tasks such as sentiment and requirement classification, but few studies compare their systems with human analysts, so it remains open whether accuracy is sufficient. We also argue that acquiring need data deserves as much attention as analyzing it. Building on these findings, we outline a design in which agents acting for users describe their needs and coordinate with agents acting for suppliers. We close with a research agenda on data availability, data acquisition, agent protocols, privacy and incentives.

Authors

Publication Details

Published
2026-11-16
DOI
https://doi.org/10.69052/jrisa.2026.1020
Primary Topic
Sentiment Analysis and Opinion Mining
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

From Passive Mining to Proactive Matching: A Survey of AI-Driven User Need Discovery and the Case for Agent-Mediated Demand-Supply Coordination

Xinglang Wang
4 links
Sentiment Analysis and Opinion Mining
article

From Passive Mining to Proactive Matching: A Survey of AI-Driven User Need Discovery and the Case for Agent-Mediated Demand-Supply Coordination

Xinglang Wang
article en

Abstract

Firms have long struggled to learn what users want before deciding what to produce. This survey reviews research published from 2024 to 2026 on using artificial intelligence (AI), in particular large language models (LLMs) and LLM-based agents, to discover user needs from user-generated content such as product reviews and social media posts. We searched Google Scholar with a bilingual English and Chinese query matrix and applied deduplication, bibliographic verification, rule-based relevance scoring and citation snowballing, which produced a verified corpus of 98 papers. Each paper was classified by technique, application domain and year. Within this corpus, LLMs are the most common technique, agent-based work is growing from a small base, and retrieval-augmented generation is rarely used. Applications are concentrated in software requirements engineering and e-commerce, while other industries remain thinly covered. We propose that this pattern reflects how much public user-written text each industry produces, and we present this as a proposition for future testing because our search design favors such industries. Reported accuracy is promising on narrow tasks such as sentiment and requirement classification, but few studies compare their systems with human analysts, so it remains open whether accuracy is sufficient. We also argue that acquiring need data deserves as much attention as analyzing it. Building on these findings, we outline a design in which agents acting for users describe their needs and coordinate with agents acting for suppliers. We close with a research agenda on data availability, data acquisition, agent protocols, privacy and incentives.

(2)
Sentiment Analysis and Opinion Mining
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon