Skip to main content

Surfient module · Radar

See which of your products AI assistants recommend — and why

The Radar runs high-intent shopper prompts against every major AI engine, segments answers by product, and shows you which SKUs are getting recommended, which are being ignored, and which are riding a competitor's wave.

  • Per-product scorecards show recommendation rate across ChatGPT, Perplexity, Claude, and Google AI Overviews — across the shopper prompts most likely to land on your store.
  • Gap reporting surfaces SKUs that should be in the consideration set but aren't — with the specific reasons the audit engine thinks they're being skipped.
  • Opportunity reports show adjacent categories where AI engines already cite you — so you know where a line extension has a head start.
See my product recommendation mapExplore the full feature list

Product Discovery Radar

Prompts matched to your catalog

94% match top

Shopper prompt · Perplexity

“best merino base layer for ski touring under $200”

Alpine Merino Base Layer 210

merino210gsmunder $200

$168

94% match

Ridge Zip-Neck Midweight

merinozip-neckski

$182

88% match

Summit Long-Sleeve Crew

crewmerinolightweight

$149

81% match

0

top match

0

2nd match

0

3rd match

The problem

Your bestseller in Shopify isn't your bestseller on ChatGPT

Revenue data tells you what converts. But AI engines shape what gets considered — and if they're recommending a different product than the one your marketing team has been pushing, the gap compounds. You keep spending ad budget on SKU A while AI quietly funnels shoppers toward SKU B. Without the radar, you never see it.

  • 43%

    difference between "top 5 products by revenue" and "top 5 products by AI recommendation rate" on the average Shopify store

    Our analysis across 120 merchants — the gap is wider than almost anyone expects going in.

  • 1 of 3

    stores has a hero product with zero AI recommendation rate in its own category

    Usually a schema or citation-readiness issue on that specific SKU. The audit engine finds it — the radar shows why it matters.

  • 2-6×

    revenue lift when merchants realign ad spend around AI-recommended SKUs

    Internal case study across 14 merchants who committed to the realignment over a quarter.

How it works

Match the prompts that matter to the SKUs you sell

The Radar is what happens when you take the Visibility Monitor's panel and pivot it by product instead of by category.

  1. Build the shopper prompt set

    For every category you sell into, Surfient proposes a set of shopper-intent prompts: "best X for Y," "cheapest X under $Z," "alternative to [competitor SKU]," "X recommended by experts." Approve the set, add yours, and save it as your store's recommendation panel.

  2. Run the panel across engines

    Every week, the full panel runs against ChatGPT, Perplexity, Claude, and Google AI Overviews. Each answer is parsed to extract every recommended product — yours and competitors'. Mention + link = strong, mention only = medium, generic category suggestion = weak.

  3. Pivot by SKU

    Instead of "how often did our store get cited" (category-level), we pivot to "how often did this specific SKU get recommended." Each product gets a scorecard: recommendation rate per engine, week-over-week trend, top-cited competitors on the same prompts, and the audit findings that might explain the delta.

  4. Surface gaps and opportunities

    Gap reports: SKUs that should be in the consideration set (matching price, specs, category) but aren't being recommended. Opportunity reports: adjacent prompts where you're already recommended — good hunting ground for a line extension. Both feed a backlog the merchandising team actually uses.

Inside the app

What you’ll see after install

Every number a Shopify merchant running Surfient Product Discovery Radar tracks in one glance — live from the Surfient admin. AI engine splits, revenue lift, and the exact state of your catalog across ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini, and Copilot.

Capabilities

What the Radar shows at each zoom level

Eight views, from store-wide down to a single SKU's answer log.

  • Store heatmap

    Every product you sell, colored by AI recommendation rate across the panel. Deep cyan = recommended often, faded = rarely, gray = never. Sort by revenue or by recommendation gap to find the biggest disconnects in seconds.

  • Per-SKU scorecard

    A product detail view with recommendation rate per engine, the specific prompts where it's recommended, the competitor SKUs that show up alongside, and the citability score from the GEO Audit Engine. One page to understand why a product is or isn't being picked.

  • Gap finder

    Filters products by criteria (price range, in stock, category, spec match) and surfaces those that match a prompt's implicit requirements but don't appear in the model's recommendation. Gap findings are severity-ranked by the prompt's estimated volume.

  • Opportunity map

    Shows adjacent prompts where your brand is already recommended generically ("Surfient-brand options include…") and flags those as places a specific SKU or line extension could convert the generic mention into a specific recommendation.

  • Competitor co-mention graph

    A graph of which competitor products your SKUs get mentioned alongside. Over time, you see who the AI treats as your substitutes and who it treats as category adjacencies — not always who you think.

  • Prompt-to-SKU matrix

    A matrix view for merchants who want granularity: every shopper prompt vs. every SKU, with recommendation status per engine. Exportable as CSV for merchandising reviews.

  • Trend timelines

    Per-SKU recommendation rate over time, annotated with the Fix Pack changes and content rewrites you shipped. Lets you correlate: "we rewrote the lead sentence on Jan 14th; recommendation rate doubled by Feb 4th."

  • Merchandising sync

    The Radar publishes a weekly "AI recommendation leaders" list to your admin. Pairs cleanly with marketing planning — you stop boosting ads for products AI isn't recommending, and start boosting the ones it is.

Customer proof

Proof

“We thought our signature roaster was our hero product. The Radar showed ChatGPT kept recommending our entry-level model instead — which had better reviews and clearer specs. We pivoted our brand story, rewrote the hero product's page, and six weeks later both were winning.”
Chiamaka Eze · Founder, Fen Coffee Roasters

2.4×

recommendation rate on realigned SKUs

Pairs well with

Keep reading across the Surfient platform

FAQ

Questions, answered straight

  • How does the Radar differ from the Visibility Monitor?

    The Visibility Monitor tracks brand-level Share of AI Voice per category — it answers "are we being cited in this category's answers?" The Radar pivots that data by product — it answers "which of our SKUs are being recommended, and which aren't?" Same underlying prompt panel, different unit of analysis. Stores that care about merchandising strategy tend to live in the Radar; stores that care about brand strategy live in the Visibility Monitor. Most customers use both.

  • Can it handle a catalog of tens of thousands of SKUs?

    Yes. We've tested with a 42,000-SKU home goods store. The per-SKU scorecard only computes for products that actually surface in AI answers (roughly 5-15% of catalog in most stores), and the heatmap aggregates long-tail SKUs into their collections. Performance stays snappy because we don't force compute you don't need.

  • What if my products don't show up in any AI answer yet?

    That's the most common starting state, and honestly a useful one. The Radar still runs the panel and shows you who's being recommended in your target prompts — your competitors. Combined with the GEO Audit Engine, you'll know the schema, citation-readiness, and content gaps specific to each of those winning pages. Close them, and recommendation rate follows.

  • Can I use the Radar to plan new product launches?

    Yes, and it's one of the highest-leverage uses. Before launching a new SKU, run a prompt panel simulating the shopper you're targeting. The prompts where you're already generically mentioned but not specifically recommended are exactly the consideration sets your new product should slot into. Ship the SKU with the schema and citation surface that match those prompts, and the Radar tracks whether it lands.

  • How often does the data refresh?

    Weekly across the full panel. High-priority prompts can be set to daily on growth + enterprise plans. Each run is timestamped and archived — no data gets overwritten, so trend timelines remain exact.

  • Can the Radar tell me why a specific SKU isn't being recommended?

    Usually yes. The Radar joins recommendation data with the GEO Audit Engine's findings, so a scorecard can say "this SKU has no Offer schema, its lead sentence is 64 words, and the product page is missing 3 of the 5 attributes the prompt asks about." That gives merchandisers a concrete fix backlog instead of a vague mystery.

Find out which SKUs AI is already hand-selling for you

The free scan runs 50 shopper-intent prompts against four major engines and gives you a ranked list of your recommended SKUs, your ignored SKUs, and the competitors winning the prompts you lose.

Run my product recommendation scanSee pricing