Skip to main content

Surfient module · Intent

Know what shoppers are actually asking AI — before your competitors do

Vertical prompt libraries refreshed weekly, plus a research panel that surfaces the specific phrasings shoppers use in ChatGPT, Claude, and Perplexity before they land on your store.

  • Prompt libraries for 22 verticals — fashion, beauty, home, food, outdoor, B2B SaaS and more — seeded from a 12,000-shopper consented research panel.
  • Weekly shopper-intent reports show which prompts are rising, falling, and emerging — so merchandising and content teams can react before a trend calcifies.
  • Every prompt in the library is tagged with intent (consideration, comparison, validation, troubleshooting, purchase) so you prioritise work by where shoppers are in the funnel.
Explore the prompt libraryPricing

Prompt Intelligence

248 tracked prompts · 67% win rate

+11 prompts

Tracked prompts · 248 total

6 / 9 cited

“best ski touring base layers”

“merino vs synthetic”

“sustainable outdoor brands”

“alpine layering system”

“warmest lightweight mid”

“patagonia alternatives”

“avalanche beacon guide”

“best ski jackets 2026”

“merino care instructions”

0

cited

0

at risk

0

missed

The problem

You can't optimise for prompts you've never seen

Shopify dashboards show what shoppers searched for on your store. AI assistants show what shoppers searched for instead of your store. The second set is invisible to every standard analytics tool — and it's where the decisions are being made.

  • 68%

    of AI shopping prompts are phrased differently than any Google search in your GSC export

    Conversational syntax, multi-part requirements, and natural-language qualifiers don't map to keyword research the way you're used to.

  • 11 days

    median half-life of a trending shopper prompt in 2026

    Down from 28 days in 2023. Conversational AI has compressed the reaction window considerably.

  • 0

    existing keyword tools that surface the prompts AI shoppers are actually using

    Google Trends, Ahrefs, Semrush — all look at the search-era surface. Prompt Intelligence is a category of one today.

How it works

Harvest, classify, refresh, deliver

Consented shoppers tell us what they're actually asking. Classifiers normalise. Weekly delivery keeps you ahead of the curve.

  1. Harvest from the consented panel

    Surfient's research panel of 12,000 shoppers across the US, UK, EU, and AU shares their AI assistant prompts when they opt in. Every week we get tens of thousands of fresh, real shopper prompts — including the half that never appear in any keyword tool because they're conversational and long-tail.

  2. Classify by vertical and intent

    An intent classifier tags every prompt with: vertical (fashion, beauty, home, food, outdoor, fitness, pet, etc.), intent stage (discovery, consideration, comparison, validation, purchase, troubleshooting), and entity set (brands mentioned, product attributes, price range, geography). The result is a structured corpus you can query.

  3. Refresh weekly, archive forever

    Every Monday, the prior week's prompts get merged into the library. Prompts trending up move to the top of their vertical; prompts trending down recede. Nothing is deleted — the full history is queryable, so you can see a trend's full lifecycle.

  4. Deliver into your workflow

    Prompt libraries flow into the Visibility Monitor (as panel candidates), the Product Discovery Radar (as recommendation probes), and the Content Engine (as prompts to optimise lead sentences for). The weekly shopper-intent digest lands in inboxes and Slack for the PMs and merchandisers who need to think about the direction, not just the day-to-day.

Inside the app

What you’ll see after install

Every number a Shopify merchant running Surfient Prompt Intelligence 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

Inside the library

It's not a static list. It's a living corpus, structured and queryable.

  • 22 vertical prompt libraries

    Each vertical has 200–800 prompts, curated from the consented panel and refreshed weekly. Verticals include apparel, shoes, beauty, home goods, home improvement, outdoor, fitness, pet, baby, food and beverage, wellness, electronics, B2B SaaS, plus more niche categories on request.

  • Intent-stage tagging

    Prompts are classified as discovery ("what's the best X"), consideration ("is X or Y better"), comparison ("X vs Y for Z"), validation ("is X worth it"), purchase ("where can I buy X"), troubleshooting ("fix my X"), and post-purchase ("how do I use my X"). Most merchants find their citation gap is heaviest on comparison and validation — exactly where conversions are decided.

  • Trend detection

    A time-series classifier identifies rising, falling, and emerging prompts. A prompt that went from 12 mentions last month to 180 this month is an emerging trend; the platform flags those automatically, so merchandising teams can build content before the trend peaks.

  • Entity extraction

    Every prompt's brands, attributes, price ranges, and qualifiers are extracted and indexed. You can query: "show me all prompts in apparel that mention sustainability AND are under $100." Useful when planning capsule collections or eco-focused campaigns.

  • Geo-tagged prompts

    Panel participants contribute from 18 countries. Prompts are geo-tagged so US vs. UK vs. EU shopper phrasing can be analysed separately — often very different, especially for anything regulated (food, supplements, cosmetics, legal products).

  • Language coverage

    English is deepest. Spanish, French, German, Italian, Portuguese, and Dutch are GA. Japanese, Korean, and Swedish are in the consented-panel pipeline for 2026. Prompts stay in their original language; translations are provided as a convenience, never as a substitute.

  • Weekly shopper-intent digest

    Every Monday, a curated digest of the prompts most relevant to your store: rising trends in your vertical, new comparison prompts mentioning competitors you track, and emerging long-tail requirements. The digest typically runs 1–2 pages and is designed to be read in under 10 minutes.

  • API access for power users

    The entire prompt corpus is queryable via the Surfient API. Build a custom dashboard, feed prompts into your own content pipeline, or export for an in-house content strategist. Rate limits are generous — we want you using the data.

Customer proof

Proof

“The shopper-intent digest surfaced a rising prompt — "gift for a dad who camps" — three weeks before Google Trends caught it. We built a content-ready gift guide on our store in four days and won the citation within a month.”
Jasper Aalto · Head of Content, Fieldline Outfitters

3 weeks

earlier signal than standard keyword tools

Pairs well with

Keep reading across the Surfient platform

FAQ

Questions, answered straight

  • How is this different from keyword research tools?

    Keyword tools show you search queries — typically 1-6 words, grouped by volume. Prompt Intelligence shows you assistant prompts — typically 10-40 words, conversational, multi-constraint, and often invisible to keyword tools because shoppers don't use them on Google. The two are complementary: keywords tell you what Google shoppers search for; prompts tell you what AI-assistant shoppers ask.

  • How big is the research panel?

    12,000 consented shoppers across 18 countries. The panel is recruited, screened, and compensated by a third-party research firm (Prolific + a regional partner for APAC). Participants opt in explicitly and can revoke at any time. We publish panel demographics in our methodology doc so you can judge representativeness for your category.

  • How does consent work?

    Panel participants opt in to share redacted AI prompts. We strip personally identifying information before storage (names, addresses, phone numbers). Prompts that reference sensitive categories (health conditions, finances) are excluded or aggressively redacted. Participants can withdraw at any time, which removes all their prior contributions from the library.

  • Can I add private prompts to my account?

    Yes. Up to 500 custom prompts on growth and enterprise. These are private to your account and never enter the public library. They flow into your Visibility Monitor panel and Content Engine briefs the same way library prompts do.

  • How often is the library refreshed?

    Weekly. The Monday digest reflects the prior week's harvest, classified and merged into the existing library. We've been running this cadence for 14 months without missing a week.

  • Can I see which prompts my competitors are winning?

    Yes — that's where Prompt Intelligence pairs with Competitor Intelligence. Prompts in your vertical get run against ChatGPT, Perplexity, Claude, and Google AI Overviews as part of your Visibility Monitor panel. For each prompt, you see which named brands got cited. Cross-referenced with Competitor Intelligence, you see who's winning what and why.

Read the prompts your shoppers are sending to ChatGPT this week

Pick a vertical. We'll share a sample of the top 100 trending prompts from the prior week's panel harvest — no account required. If the signal is useful, keep reading the digest every Monday.

See this week's promptsSee pricing