LinkedIn Ad Intelligence¶

Capability Slug: linkedin_ad_intelligence Plan Tier: Command Credit Cost: 10 credits per analysis Category: Competitive Intelligence Data Source: LinkedIn Ad Library API (live)

Renamed from LinkedIn Insights / Graph Analysis

This capability was previously documented as LinkedIn Insights (slug linkedin_graph_analysis) and described five modes: Post Performance, Audience Insights, Relationship Graph, ICP Alignment, and Content-to-Marketplace Correlation. Those modes required LinkedIn API access tiers that Vellocity does not have (Member Data Portability, Community Management, Marketing Developer Platform), so they shipped on user-pasted data rather than live LinkedIn data. We dropped the five unsupported modes and renamed the capability to match what it actually does — live competitor ad teardown via LinkedIn's Ad Library API. If you have automations or agents that reference linkedin_graph_analysis, the rename migration rewrites them to linkedin_ad_intelligence automatically.

Overview¶

LinkedIn Ad Intelligence lets AWS Marketplace partners reverse-engineer competitor demand-generation strategy without a single screenshot. The capability pulls live ads directly from LinkedIn's Ad Library API for any competitor or keyword, then uses Claude to extract the messaging, value props, formats, and positioning gaps that infrastructure tools can't see.

The problem it solves

Marketplace partners often discover competitor positioning the hard way — losing deals to messaging they didn't know was running. Manual ad research means rotating through LinkedIn ad library tabs, screenshotting individual ads, and trying to spot patterns by eye. It's slow, inconsistent, and biased toward whatever you happen to see.

What Vellocity does

Queries the LinkedIn Ad Library API for one or more advertisers (or a keyword)
Returns up to 25 ads per advertiser with full creative metadata
Sends the ad set to Claude with your brand context for a structured teardown
Surfaces themes, formats, value props, and gaps competitors are not covering — your positioning opportunities

What You Provide¶

Parameter	Required	Description
`analysis_type`	Yes	Currently `competitor_ads` (single supported mode)
`competitors`	One of these two	Array of advertiser names (e.g., `["Datadog", "Splunk"]`)
`keyword`	One of these two	Search term (e.g., `"observability"`, `"finops"`)
Brand context	Optional	Pulled from your company profile if present — improves the teardown

You must supply at least one of competitors or keyword. Combining them filters to ads from those advertisers that match the keyword.

What You Get Back¶

A structured JSON object containing:

insights — messaging themes, format patterns, positioning observations
recommendations — content themes to differentiate, messaging gaps to exploit, ad format suggestions, AWS Marketplace positioning angles
competitor_summary — per-competitor breakdown (ad count, dominant themes, top CTAs)
ad_library_data — raw ads from the Ad Library so you can drill into specific creatives
summary — short executive read

Example output (abbreviated)

{
  "insights": [
    "All three competitors lead with cost-reduction messaging (40-60% savings claims)",
    "Video ads dominate (72%) — single-image ads are under-used",
    "None of the competitors mention AWS Marketplace Private Offers in their copy"
  ],
  "recommendations": [
    "Lead with 'time-to-value' rather than cost-savings — differentiated angle",
    "Test single-image carousel ads with stat-led headlines",
    "Build a 'buy through Marketplace' campaign — uncontested messaging space"
  ],
  "data": {
    "competitor_summary": {
      "Datadog": { "ad_count": 18, "top_theme": "unified observability" },
      "Splunk": { "ad_count": 12, "top_theme": "AI-driven alerts" }
    }
  },
  "summary": "Field is crowded on cost; Marketplace-buying motion is wide open."
}

When to Use It¶

Scenario	Why this capability fits
Pre-launch positioning	Confirm you're not parroting what every competitor already says
Quarterly GTM review	Track when competitors shift messaging (re:Invent, funding events)
AWS Marketplace listing copy	Ground your hero copy in real competitor gaps, not internal opinions
New ICP entry	See what messaging is winning attention in a new vertical before you spend

It is not a replacement for:

Native LinkedIn analytics on your own posts (LinkedIn's dashboard does that)
Pipeline attribution (use Pipeline Influence Tracking — deal_influence_tracking)
Audience demographics (LinkedIn's own audience tools)

Getting Started¶

Connect LinkedIn — Integrations → Social Media → Connect LinkedIn (publishing scope is sufficient; analytics scope works too)
Open Marketing & Social → LinkedIn Ad Intelligence
Enter competitors and/or a keyword — start with 2-3 direct competitors
Run the analysis — typically returns in under 30 seconds
Review the teardown — focus on the recommendations and competitor_summary.gaps sections

API Access¶

LinkedIn Ad Intelligence is exposed as a capability through the agent runtime and the Partner API. To call it via the Partner API, scope a key to the linkedin_ad_intelligence capability and POST analysis requests through the agent execution endpoint. See Partner API Capabilities and Authentication & API Keys.

Frequently Asked Questions¶

Q: Can I analyze competitors that don't run LinkedIn ads? A: No. This capability reads LinkedIn's Ad Library — if a competitor doesn't advertise on LinkedIn, there's nothing to analyze. Use Competitor Analysis (competitor_analysis) for non-ad competitive intelligence.

Q: How many ads does it pull per competitor? A: Up to 25 per advertiser, per query. For competitors running large campaigns, this returns the most recent slice — paginate the underlying Ad Library service to go deeper.

Q: Will Vellocity post on my behalf? A: No. The capability is read-only against the Ad Library. To publish ads or organic content, use Social Media Publishing (publish_to_social_media).

Q: What happened to the relationship-graph and warm-path features? A: They were removed. Building a real connection-graph product requires LinkedIn API access tiers Vellocity doesn't have. Rather than ship prompt-engineered estimates and call it "graph analysis," we cut the feature. Warm-path discovery is not currently on the roadmap unless LinkedIn opens up the relevant data access.

Q: Is content-to-marketplace correlation gone too? A: As a LinkedIn-specific feature, yes. Pipeline Influence Tracking (deal_influence_tracking) handles content → pipeline attribution across all sources, including LinkedIn UTMs.

Credit Cost Reference¶

Operation	Credits
Competitor ad teardown (one or more advertisers, optional keyword filter)	10
Keyword-only ad search	10

Cached query results (same competitors + keyword within 1 hour) do not re-bill — only fresh API calls consume credits.

Last Updated: 2026-04-27 Capability Version: 2.0.0 (renamed from linkedin_graph_analysis 1.x)