F&T Intelligence · AI Search Report
fever-tree.com
AI Search dimension · April 2026 · Delivered by Friction & Toil
AI Visibility Maturity Curve
Emerging
AI Share of Voice (5 / 18 cells): Google 2/6 · Bing 1/6* · Brave 2/6  *Bing forced Unlikely on 4 of 6 queries (natural-language tokenisation failure)
Add Product and FAQ structured data to every fever-tree.com product page. This is the single fix most likely to move Fever-Tree from Emerging to Cited (6/18 minimum) for "compare premium tonic waters" and "best cocktail mixer" searches, where AI engines currently extract specs from Difford's Guide and Amazon instead of your own site.
✓ Blocked
Invisible
Emerging

This snapshots your structural readiness across Google, Bing (the index that powers ChatGPT search and browse mode), and Brave (the index that powers Anthropic Claude search and Perplexity). For ongoing drift tracking, week by week, of what AI is actually saying about you in the wild, see the £150 / mo Pulse.

Six target queries: who surfaces, and why
Query Structural verdict Bing Brave Who surfaces instead
"Fever-Tree tonic water" Likely Cited (own surfaces only)
"Fever-Tree mixer range" Possible Bing tokenised to NHS / fever results
"best premium tonic water for gin UK" Unlikely N/A Telegraph, Good Housekeeping, Craft Gin Club
"premium cocktail mixers comparison" Unlikely N/A Amazon, goodtimein, Funkin (Bing returned NS&I Premium Bonds)
"best ginger beer for Moscow mule" Unlikely N/A Slightly Pretentious, Savory Reviews, Feast & West
"what is the best tonic water?" Unlikely N/A Tasting Table, Good Housekeeping, Bon Appétit, Liquor.com

Structural verdict: assessed against documented AI citation factors based on structural and content signals (Likely Cited / Possible / Unlikely). Bing & Brave: live-tested May 2026 via WebFetch on bing.com and search.brave.com (citation = fever-tree.com in top 5 organic OR named in AI summary box). Bing forced Unlikely on 4 of 6 queries due to natural-language tokenisation failure (Bing returned dictionary / NS&I Premium Bonds / NHS results for unquoted natural-language queries, a known engine-specific failure mode in this niche, not a Fever-Tree gap). Live prompt testing against ChatGPT, Perplexity, Gemini, and Claude is available as a retainer add-on.

The publishers that control your AI reputation in Premium Beverages / Cocktail Mixers
Host AI citation frequency Your presence
diffordsguide.com High Active
tastingtable.com High Active
masterofmalt.com High Active
imbibemagazine.com Medium Mentioned
craftginclub.co.uk Medium Mentioned

These are the external sources AI engines pull from when answering category queries in your sector. Presence on high-frequency hosts is the fastest lever to move up the Maturity Curve. If you work with a PR agency, share this table with them: it's their placement brief.

Can AI compare you head-to-head when asked?

Bake-Off Absent (0 / 3)

No pricing is visible as HTML text anywhere on the site. Purchasing redirects entirely to third-party retail partners (Waitrose, Amazon, Ocado). Product features are presented in marketing prose rather than structured tables or lists. There are no competitor comparison claims backed by specific numbers or specifications. When AI engines answer "compare Fever-Tree vs Q Mixers vs Double Dutch," they pull every spec from Difford's Guide profiles and Amazon listings, because fever-tree.com provides no machine-readable product data of its own.

Can AI engines read your site?

Fully Open

No AI crawlers are blocked. GPTBot, ClaudeBot, PerplexityBot, GoogleOther, CCBot, anthropic-ai, OAI-SearchBot all have unrestricted access to fever-tree.com. The only Disallow rule is /admin. This is the correct configuration. Crawler access is not the bottleneck here.

Your next move
Add product schema to all mixer pages
Fever-Tree's product pages contain zero structured data: no schema.org/Product markup, no ItemList, no FAQ schema. When AI engines are asked to compare premium tonic waters or recommend a Moscow Mule mixer, they extract ingredients, formats, and specs from Difford's Guide and Amazon listings rather than from fever-tree.com itself. Adding Product schema (name, description, ingredients as structured properties, available pack sizes) to every product page makes each SKU directly extractable by AI citation engines. This is a single sprint dev task with no content rewrites required, and it directly targets the primary citation gap across all 6 audited queries.

Get the full five-dimension report

This report covers AI Search only. The full Web Intelligence Report adds Design, Brand, Competitive position, and Growth, with a £-denominated business impact model and three competitor deep-dives.

Your £250 credit toward the full report:

AI-SEARCH-UPGRADE

View full report options →