1. What Are AI Insights?
AI Insights are thematic, data-driven cross-sections of the Tasting Intelligence database. Each insight explains in plain words why a group of drinks, sharing a cask, a finish, a method or a country, tastes the way it does, and what that means when you choose or taste a bottle.
Think of an AI Insight as a curated lens on our data: instead of browsing hundreds of individual expressions, an insight surfaces the hidden patterns and correlations that connect them. It answers questions like: "Which Speyside single malts with sherry cask finishes share the most coherent fruity-floral profile?" or "What do the top-rated peated expressions have in common beyond smokiness?"
Example: At a glance
- Drinks
- 12
- Ratings
- 86
- Average rating
- 8.1
- Data strength
- Moderate
2. How to Read an AI Insight
Every AI Insight puts the answer first. The summary at the top tells you what it is about in one or two sentences; read further only for the parts that interest you.
Key takeaways
Three short points with what stands out in this group of drinks. If you only read these, you know the story.
The why, one question at a time
Each takeaway gets its own short section, with a plain question as its heading. Spirits terms are explained the first time they come up. When the data is thin, the text says so.
Try these
Two to four bottles that show the pattern well, each with one sentence on why. Every named bottle comes from our database.
Under the takeaways, two charts show the data behind them: how often each flavor group comes up compared with similar drinks, and which casks and finishes the bottles have. They use the same data the text was written from.
3. How Alike They Taste
How alike they taste shows how clearly the drinks in an insight share a flavor group. It combines two things: how many of the drinks have that group, and how strongly it comes through. The sidebar lists the four strongest groups.
A flavor band is a group of related taste notes, for example "Fruity & Floral" (apple, pear, citrus, rose) or "Smoky & Peaty" (bonfire, iodine, medicinal).
High Coherence (75–100%)
Stable, harmonious flavor clusterMost expressions share this band, and clearly. If you enjoy one bottle in this insight, you are likely to enjoy the others. A good starting point when you want more of a style you already love.
Lower Coherence (below 50%)
Diverse or exploratory clusterFew expressions share this band, or only faintly. That is not a flaw: it means the band is an accent, not the core (for example smoke in a mainly fruity insight), or the group is more diverse.
How to Use This for Bottle Discovery
Look at the top two or three bands: they define the core of the insight. If they match your taste, the bottles under Try these are a safe place to start. For something more adventurous, pick an insight where your favorite band scores in the middle (50–75%): the bottles share your notes but add some surprise.
4. The Numbers at a Glance
The sidebar of every AI Insight sums up what it rests on. The numbers are updated with each data sync (about once or twice a month), so they can lag slightly behind the latest reviews.
Drinks, ratings and average
How many drinks belong to the insight, how many ratings they have together and their average rating, in which every rating counts once.
Data strength
How reliable the picture is. It grows with more reviews, more reviews from our community and raters who agree with each other. Strong means the pattern is stable; Moderate and Limited mean: read it as a direction, not a verdict.
5. Reading the Rest of the Page
- Charts: under the takeaways, how often each flavor group comes up compared with similar drinks, and which casks and finishes the bottles have.
- Try these: a short selection of bottles from the group, each with the reason it is a good way to taste the pattern. Every name links to its drink page.
6. Under the Hood
The models and software behind this page, and when they run.
-
Gemini 3.8 Flash Claude Sonnet 5.5 Gemini 3.5 Flash Lite Gemini 3.1 Pro Preview
Writing AI insights: First an evidence pack is built from our database, without AI. Gemini turns it into a research brief, Claude writes the article in plain language, and a lighter Gemini model writes the one-line summary. When a model is unreachable, Gemini Pro takes over. An admin reviews every insight before it is published.
When: Once per insight, started by an admin.
-
Gemini 3 Pro Image
Images: Creates an illustration for a drink without a photo (with a watermark) and the header image of an AI insight.
When: When a drink or insight is created without an image.
-
Gemini 3.8 Flash Claude Sonnet 5.5 Google Cloud Translation
Site translations: Translates the site and its content into Dutch, French, Spanish and German. Gemini goes first; when it fails, Claude and then Google Cloud Translation take over. Every translation is stored and reused.
When: Ahead of a release, or the first time a page is visited in a language.
About the AI Analysis
The analysis is written with AI on Vertex AI from Google Cloud, but only from a fixed set of facts calculated from our database: how the group differs from similar drinks, its ratings and its bottles. Findings without supporting data are dropped, and bottles can only come from the database. Bot reviews do not count as organic reviews. The AI provides the interpretation; the data provides the foundation.