Methodology
How Synopsis reads reviews.
Synopsis turns public hospitality feedback into short, useful place profiles. This page explains what we collect, how the AI analysis is generated, what can change, and how to ask for a correction.
Last updated - June 26, 2026
The short version.
Synopsis analyzes public review patterns from external sources where those sources are available for a place.
AI summaries are based on repeated review signals, source coverage, and visible guest feedback, not paid placement.
Scores and summaries are directional. They can change when source data, review volume, venue details, or analysis models change.
Business owners can request corrections or claim their profile when public data is incomplete or outdated.
01
What data sources does Synopsis use?
Synopsis works with public place and review information where it is available. Typical sources include external review platforms, reservation profiles, and public place-data providers for restaurants, hotels, bars, and other hospitality venues.
Not every place has every source. Some venues have broad review coverage but no reservation-side profile, some have travel reviews but no booking-side data, and some sources may be unavailable during a refresh. When a source is missing, Synopsis continues with the available evidence instead of pretending the signal exists.
- Place identity signals such as name, address, category, and public source URLs.
- Review text, ratings, dates, source labels, and visible reviewer context where available.
- Venue photos and metadata when they help identify or present a public place.
02
How is the AI analysis generated?
Synopsis groups guest feedback into recurring themes: food, service, ambience, value, noise, access, dishes, and other signals that appear often enough to matter. The system looks for repeated patterns across sources rather than treating one loud review as the whole story.
The AI then writes a concise synopsis in plain language. The goal is not to replace the original reviews; it is to show the consistent shape of what people say so readers and operators can decide what to inspect next.
- Repeated praise and criticism are weighted more strongly than isolated comments.
- Aspect scores are directional summaries of sentiment, not official ratings from the source platforms.
- Partner tools add operational views, such as trends, reply support, and profile editing, after the business is verified.
03
How fresh is the information?
Public place pages and partner dashboards can update when a new analysis is run, source data is refreshed, or a verified partner requests an update. Some partner features use scheduled review updates and broader dashboard refreshes.
Freshness depends on source availability, review volume, and whether a venue has enough public information to support a reliable read. When a place has little data, Synopsis may show a lighter profile or wait until more evidence is available.
- Newer reviews can shift the synopsis when they repeat a meaningful pattern.
- Older reviews can still matter when they describe a persistent issue or strength.
- Source outages, removed listings, or platform changes can delay a refresh.
04
What are the limitations?
Review data is imperfect. It can be sparse, biased, duplicated, translated, sarcastic, outdated, or influenced by unusual one-off events. AI can help read scale, but it cannot personally visit the venue or verify every claim in every review.
Synopsis is best used as a clear starting point, not as the final authority on a place. We aim to make uncertainty visible by showing source coverage, recent signals, and the themes behind a summary whenever possible.
- A venue with few reviews may have a less stable synopsis than a venue with broad, recent coverage.
- AI summaries may miss nuance in slang, multilingual reviews, sarcasm, or local context.
- Scores should be read as a practical signal, not as a guarantee of quality or safety.
05
How can businesses request corrections?
If you own or manage a listed place and something is wrong, contact us with the place name, address, source links, and the correction you believe is needed. We review correction requests manually before changing public-facing information.
Verified partners can also claim their profile, add first-party context, and keep operational details current without changing the independent review analysis.
- Correction requests: info@evolvx.it
- Privacy or data rights requests use the process described in the privacy and terms page.
- We do not sell rankings or remove legitimate criticism in exchange for payment.
Related transparency pages.