Turning reviews into an intent graph: how platforms infer what guests really want
Machines do not just read sentiment in reviews, they reconstruct unstated needs — and that drives which guests each property gets matched to.
Key numbers
- Method: build a user-specific intent knowledge graph from each review, align it with a global hotel knowledge graph through structure-aware semantic smoothing, then let an LLM jointly predict aspect ratings and generate reverse user-intent statements.
- Comparison: the framework consistently outperforms strong LLM and retrieval-based baselines on both classification and intent generation.
- Validation data: online travel reviews from TripAdvisor.
- The authors argue structure-aware knowledge alignment enables scalable, explainable intent inference.
Method and sample
arXiv paper “Mind the Gap: A Dual Knowledge Graph Framework for Unified Multi-task User Intent Inference” (2026-08-07): the DKG-MTI dual knowledge graph framework infers user intent from online travel reviews, benchmarked against strong LLM and retrieval baselines on TripAdvisor reviews for classification and intent generation.
What it means for brands
Platforms are already structuring reviews into intents and using them to decide which guests see which property. To be matched with the right guests, your reviews and pages must explicitly carry the dimensions those guests care about — location, quiet, family-friendly, pet-friendly.
How SourceSight handles it
SourceSight groups prompts by how customers actually ask (is it right for me, nearby options, pricing, policy questions) to find the question types where you are absent, and the off-site list prioritises review and local directory sources.
Limits
The study uses TripAdvisor data and task-specific settings, so it is a methods comparison rather than a description of platform ranking rules or of every market's data structure.
References
Run it on your own project
What the same prompts return across 10 engines is only knowable by running them.
Related research
Cross-category recommendations: +36% add-to-cart on item pages, +15% on cart pages
Recommending general merchandise to grocery shoppers — a milk frother with the milk — has long been hard; LLM retrieval plus real-time cart context moved production A/B metrics.
Industries · 2026-10-07Home services: 22% of homeowners already use AI to find a provider, and 80% of businesses don't know how to be found
The local services funnel is moving upstream: before a homeowner opens a website or calls, AI has already synthesised reviews, reputation and service descriptions into one answer.