Skip to content

Marketplace Awareness Engine

Brand name: Vellocity Signal™ Capability Slug: marketplace_awareness Plan Tier: Command Credit Cost: 10 / 12 / 15 / 18 / 20 / 35 credits per run (varies by analysis type) Category: Intelligence & Attribution Default mode: gtm_insights Pairs with: AWS Innovation Monitor

Overview

Vellocity Signal turns a single AWS Marketplace article — an announcement, a partner story, an industry signal — into seven different GTM artifacts. Same article, seven angles: an executive read, recommended actions, listing-copy suggestions, co-sell opportunity flags, ready-to-publish social posts, and a partner outreach email.

The capability is article-level, not corpus-level. If you're scanning many feeds to find the article worth digging into, that's AWS Innovation Monitor. Once you've identified the article, Vellocity Signal extracts everything from it.

How it relates to AWS Innovation Monitor

Capability Scope Best for
AWS Innovation Monitor (aws_innovation_monitor) Corpus-level — scans many feeds, scores updates against partner profiles "Which AWS announcement matters this week?"
Vellocity Signal (marketplace_awareness) Article-level — deep multi-angle analysis on one article "Now extract everything from that announcement."

Run them in sequence: Innovation Monitor surfaces a relevant update; Vellocity Signal turns that one update into a launch-ready GTM kit. Or run Vellocity Signal directly on any article URL when you know what you want to analyze.

The seven analysis types

Type Credits Output focus
gtm_insights (default) 15 Executive summary, seller impact, market implications, strategic considerations, urgency level, affected categories
cta_recommendations 12 Immediate actions, short-term strategy, long-term positioning, risk mitigation, competitive advantage framing
listing_optimization 18 should_update flag + priority, title/description suggestions, keyword opportunities, messaging adjustments, feature highlights, compliance considerations
cosell_opportunities 20 opportunity_detected flag + type, integration signals, partnership trends, AWS program mentions, target partners, recommended outreach, ACE eligibility impact
social_post 10 LinkedIn post + Twitter post (each with text, hashtags, character count), shared talking points, CTA
partner_email 12 Subject line, body, CTA, tone, urgency, follow-up suggestion
comprehensive 35 All six modes combined in one response (~60% of the per-mode total of 87 credits)

All modes run on Claude Sonnet 4.5 Global. Token caps and temperatures vary: analytical modes use 0.3-0.4 temperature; creative modes (social posts, email) use 0.5-0.6.

Three ways to pass an article

The capability is flexible about how it gets the article content:

Method Use when
article_data (dict with title, content/description, optional link) You already have the article content — passing it directly avoids any fetch
rss_feed_url (URL string) You want the capability to pull the first article from a feed
title + rss_feed (both strings) Existing UI pattern — the user picked an article from a rendered RSS list and you want to look it up by title within that feed

If multiple inputs are provided, article_data wins. If none resolve to a usable article, the capability throws.

What gtm_insights returns

{
  "success": true,
  "analysis_type": "gtm_insights",
  "article_title": "...",
  "article_url": "...",
  "credits_used": 15,
  "gtm_insights": {
    "executive_summary": "...",
    "seller_impact": ["...", "..."],
    "market_implications": ["...", "..."],
    "strategic_considerations": ["...", "..."],
    "urgency_level": "high|medium|low",
    "affected_categories": ["...", "..."]
  },
  "_schema": { /* MarketplaceAwarenessSchema-typed projection */ }
}

The _schema block is a typed projection of the result keyed by MarketplaceAwarenessSchema — useful for dashboards or automations that need a stable shape across analysis types.

What cta_recommendations returns

{
  "cta_recommendations": {
    "immediate_actions": ["...", "..."],
    "short_term_strategy": ["...", "..."],
    "long_term_positioning": ["...", "..."],
    "risk_mitigation": ["...", "..."],
    "competitive_advantage": "..."
  }
}

Three time-horizon buckets plus risk and competitive framing.

What listing_optimization returns

{
  "listing_optimizations": {
    "should_update": true,
    "priority": "high|medium|low",
    "title_suggestions": ["...", "..."],
    "description_updates": ["...", "..."],
    "keyword_opportunities": ["...", "..."],
    "messaging_adjustments": ["...", "..."],
    "feature_highlights": ["...", "..."],
    "compliance_considerations": ["...", "..."]
  }
}

The should_update boolean is the headline read — if false, you can skip the rest of the suggestions for this article. If true, priority and the suggestion arrays tell you what to change.

What cosell_opportunities returns

{
  "cosell_opportunities": {
    "opportunity_detected": true,
    "opportunity_type": "integration|partnership|expansion|...",
    "integration_signals": ["...", "..."],
    "partnership_trends": ["...", "..."],
    "aws_program_mentions": ["AWS ISV Accelerate", "ACE", "..."],
    "target_partners": ["...", "..."],
    "recommended_outreach": ["...", "..."],
    "ace_eligibility_impact": "..."
  }
}

Like listing_optimization, the headline is a boolean: opportunity_detected. If true, the supporting fields explain what kind of opportunity and how to act.

What social_post returns

{
  "social_posts": {
    "linkedin": {
      "post": "...",
      "hashtags": ["...", "..."],
      "character_count": 0
    },
    "twitter": {
      "post": "...",
      "hashtags": ["...", "..."],
      "character_count": 0
    },
    "talking_points": ["...", "..."],
    "call_to_action": "..."
  }
}

Both posts are returned formatted for the platform's conventions. character_count is computed via mb_strlen so multi-byte characters are counted correctly (relevant for Twitter's 280-character limit).

What partner_email returns

{
  "partner_email": {
    "subject_line": "...",
    "body": "...",
    "call_to_action": "...",
    "tone": "professional|consultative|urgent|...",
    "urgency": "high|medium|low",
    "follow_up_suggestion": "..."
  }
}

A single email — subject + body + CTA — plus tone and urgency tags so you can route it through the right outreach channel.

What comprehensive returns

A single response containing all six analysis blocks at the top level. Use this when you want everything from an article in one call (and you're willing to pay 35 credits instead of running modes individually).

{
  "success": true,
  "analysis_type": "comprehensive",
  "credits_used": 35,
  "gtm_insights": { /* same as gtm_insights mode */ },
  "cta_recommendations": { /* same */ },
  "listing_optimizations": { /* same */ },
  "cosell_opportunities": { /* same */ },
  "social_posts": { /* same */ },
  "partner_email": { /* same */ }
}

Required and optional parameters

Parameter Required Description
analysis_type Yes One of gtm_insights, cta_recommendations, listing_optimization, cosell_opportunities, social_post, partner_email, comprehensive
article_data One of these three required Dict with title and content/description
rss_feed_url URL to fetch first article from
title + rss_feed Title to look up within a known feed (UI-list pattern)

Brand-voice context is pulled automatically from the company profile via BrandContextBuilder.

When to use it

Scenario Recommended mode
AWS just made an announcement — what does it mean for our motion? gtm_insights
Quick read on whether to update our listing listing_optimization
Newsletter / Slack-channel post for the team about an announcement gtm_insights (executive_summary + seller_impact)
Co-sell signal scan on an industry article cosell_opportunities
Day-of social posts and partner email about an AWS news drop comprehensive

It is not:

  • A feed scanner. Use AWS Innovation Monitor for corpus-level scoring across many articles.
  • A general-purpose blog generator. The output is article-derivative — every block is grounded in the input article. For free-form blog generation see Generate Text Content (generate_text) or AWS Blog Co-Author (aws_blog_co_author).
  • A scoring or prioritization engine. The priority and urgency fields are AI-assessed for the article in isolation, not weighted against your partner profile.

API access

Exposed through the agent runtime as marketplace_awareness. Partner API keys scoped with manage_content and view_marketing_analytics plus the agent capability slug can run all seven analysis types programmatically. See Partner API Capabilities.

See also

Capability slug: marketplace_awareness · Brand: Vellocity Signal™ · Handler: MarketplaceAwarenessCapability · Model: Claude Sonnet 4.5 Global · Schema: MarketplaceAwarenessSchema