Agent Discoverability Score¶

Measure and improve how AWS Marketplace agents and AI assistants surface your product to buyers.

Why this matters¶

AWS Marketplace buyers increasingly start with an AI-mediated query — through Marketplace's own agent surfaces, ChatGPT, Claude, or Perplexity — before clicking into a listing page. Whether your product shows up in that response depends on signals AI grounding pipelines actually look at: cross-source consistency between your listing, your website, and third-party validation.

The Agent Discoverability Score gives you a single number that reflects how well-aligned and discoverable those signals are right now, plus specific recommendations to improve.

Score components¶

Your overall score is a weighted combination of three signals available today:

Listing Quality¶

How well your listing content is optimized for both human readers and agent grounding:

Title optimization — Keywords, length, clarity
Description completeness — Short and long descriptions
Feature highlights — Number and quality of highlight bullets
Media assets — Images, videos, logos
Category alignment — Correct category and subcategory

Backlink Authority¶

External signals that boost your listing's credibility:

Referring domains — Unique domains linking to your listing
Link quality — Authority of linking sites
Anchor text — Relevance of link text
Growth trend — Are you gaining or losing links

AI Visibility¶

How AI models reference your product when buyers ask requirement-shaped questions:

LLM mentions — Frequency of AI model references in your category
Context accuracy — Whether AI describes your product correctly
Category presence — Mentioned in relevant solution categories
Competitive positioning — How you compare to competitors in AI responses

Score interpretation¶

Score	Grade	Meaning
90-100	A+	Excellent — listing is highly discoverable
80-89	A	Strong — minor improvements possible
70-79	B	Good — several optimization opportunities
60-69	C	Fair — significant room for improvement
Below 60	D	Needs attention — multiple issues identified

Weakness detection¶

The scoring engine identifies specific issues and recommended actions:

Weakness	Impact	Recommended Action
Missing keywords in title	High	Add target keywords to listing title
Short description	Medium	Expand to 150+ characters with keywords
Few highlights	Medium	Add 5+ feature highlights
No product images	High	Add at least 3 product screenshots
Low backlink count	Medium	Create content that earns links
AI inaccuracy	Low	Update public content for AI training

Getting started¶

Navigate to Marketplace Metrics → Agent Discoverability
Select the listing to analyze
Run the analysis (takes 30-60 seconds)
Review your score and breakdown
Follow the recommendations to improve

Best practices¶

Focus on high-impact items first

Start with weaknesses marked "High" impact — title optimization and media assets typically yield the biggest improvements.

Re-score after changes

Run a fresh analysis after any listing update to confirm the change moved the score.

Monthly cadence

Schedule monthly score checks to catch drift early — search rankings shift weekly even when your listing doesn't.

Plan availability¶

Feature	Starter	Accelerate	Command
Basic listing analysis	Yes	Yes	Yes
Full Agent Discoverability Score	—	—	Full
AI Visibility analysis	—	—	Full
Competitor comparison	—	—	Full

Credit cost formula (`marketplace_listing_seo`)¶

The credit cost is computed by MarketplaceListingSEOCapability::getEstimatedCredits() — not a flat 25.

total = 15 (base)
      + 2 × competitor_count
      + 10 if include_ai_recommendations is true (default ON)

Default cost with no competitors and AI recommendations enabled: 25 credits (15 + 10). Common configurations:

Config	Cost
No competitors, no AI recommendations	15
No competitors, AI recommendations on	25
3 competitors, AI recommendations on	31 (15 + 6 + 10)
5 competitors, AI recommendations on	35 (15 + 10 + 10)

Required parameters: partner_domain, brand_keyword. Optional: competitors (array), include_ai_recommendations (default true), force_refresh (default false).

The internal capability slug remains marketplace_listing_seo for API and integration compatibility; the user-facing name and surface is Agent Discoverability Score.