# Product Recommendations with BuyScout® AI Store Agent

Turn a shopper's needs and constraints into a focused shortlist drawn from the merchant's current catalog.

## How does BuyScout® recommend products?

BuyScout® AI Store Agent uses the shopper's stated needs alongside current catalog information and available cart context. It can ask a short follow-up, identify valid options, and explain why each option fits without declaring an unsupported universal winner.

## Shopping decisions that need conversation

Recommendations are most useful when the shopper can describe the goal but cannot translate it into a product filter.

1. ### Several requirements at once
   A shopper needs a product that meets a use case, compatibility requirement, budget, and preference that no single category page explains.

2. ### Similar products with meaningful trade-offs
   A shopper sees several plausible options and needs help understanding which documented differences matter for the intended use.

3. ### The shopper knows the goal, not the category
   A shopper describes the goal but not the category. BuyScout® uses their constraints to build a shortlist from catalog evidence.

## From request to explainable shortlist

BuyScout® preserves the shopper's constraints before ranking preferences or offering alternatives.

1. ### Clarify the decision
   BuyScout® identifies the goal and asks only for missing details that would materially change which products qualify.

2. ### Check current catalog evidence
   The AI Store Agent matches the request against documented product, variant, availability, price, and compatibility information available to it.

3. ### Explain the shortlist
   Each recommendation names the relevant match and trade-off so the shopper can correct a preference or choose confidently.

## A shopper needs a complete laptop setup

**Shopper:**  Which dock works with my ThinkBook 21KG and one 4K display?

BuyScout®: Do you also need the dock to charge the laptop, and if so, how much power?

**Shopper:** Yes, up to 65 W.

BuyScout®: Checking model 21KG against the dock specifications

BuyScout®: The Portwell Dock 12 USB-C Dock supports model 21KG, one 4K display at 60 Hz, and up to 65 W charging. Add the SignalCore HDMI 2.0 Cable for the documented display connection.

-

Portwell Dock 12 USB-C Dock $129 Add to cart

-

SignalCore HDMI 2.0 Cable $18 Add to cart

## Catalog foundations for better recommendations

Clear merchandising data matters more than persuasive wording when a recommendation turns on a specific fact.

1. ### Decision-ready attributes
   BuyScout® reads public materials, dimensions, ingredients, intended use, compatibility, and care details from the relevant Shopify product page, plus any supplementary guidance in BuyScout® Dashboard.

2. ### Correct variant scope
   BuyScout® reads price, availability, color, size, and other variant-specific facts from the correct option in Shopify product and variant data.

3. ### Recommendation rules
   BuyScout® recommends from the current catalog automatically. Merchants can add hard exclusions, category priorities, buying questions, or no-match guidance in BuyScout® Dashboard.

## How Product Recommendations work automatically

BuyScout® can build and explain a relevant shortlist from the shopper's request and current catalog without separate skill configuration.

1. ### Understand the request
   BuyScout® identifies the shopper's goal, stated preferences, budget, compatibility needs, and other decision constraints from the conversation.

2. ### Search the live catalog
   BuyScout® automatically checks current Shopify products, variants, prices, availability, product pages, and other relevant website information.

3. ### Explain and refine the shortlist
   BuyScout® explains why each option fits, asks only for decisive missing details, and updates the shortlist when the shopper changes a preference.

4. ### Optionally add merchant guidance
   Merchants can optionally add store-specific buying rules, tailor recommendations, and test representative decisions in BuyScout® Dashboard.

## Recommendation boundaries

A useful recommendation is traceable to store evidence and remains easy for the shopper to challenge or refine.

- **No universal best product** The strongest option depends on the shopper's criteria. BuyScout® explains the fit without asserting an unsupported winner.
- **No mixing between options** BuyScout® does not combine the price, features, or availability of one size, color, or configuration with those of another.
- **No forced match** When the catalog lacks a valid option, BuyScout® says so or sends the shopper's request for merchant review when an exception may apply.

## Frequently asked questions

- Can BuyScout® recommend from the live catalog?
  BuyScout® automatically uses current catalog information from the connected store. Merchants can test representative products, variants, availability states, and no-match requests to confirm the shopper experience.
- What if no product satisfies the request?
  BuyScout® explains that no verified match is available, preserves the shopper's constraints, and invites a useful revision rather than recommending an invalid product.
- Can a shopper revise the shortlist after correcting a preference?
  BuyScout® reevaluates the shortlist against the revised requirement and current catalog evidence, then explains what changed instead of preserving an option that no longer qualifies.
- Does BuyScout® use shopper browsing as a hard requirement?
  BuyScout® uses permitted context to keep a response relevant while treating the shopper's stated constraints as controlling. Behavioral clues never override a direct answer.
- Can BuyScout® compare variants within one product?
  BuyScout® compares variants using price, availability, specifications, and other differences attached to the correct options. Facts remain scoped to the variant they describe.
- Can merchant priorities override required compatibility?
  BuyScout® treats compatibility on the product detail page or in BuyScout® Dashboard as a hard constraint. Ranking preferences apply only among options that still satisfy those constraints.
