Home
Guides
  1. schema
  2. Company Signals & Insights

Patent Filings

Track new patent filings to identify companies investing in innovation within your product category.

Patent Filings reveal which companies are actively investing in R&D — and what specific technology areas they're prioritizing — before those investments become public products or press releases.

We monitor USPTO assignment records across 2M+ companies, matching patent assignees to company domains. When a company files multiple patents in a technology cluster within a rolling window, we emit a signal with patent count, technology categories extracted from patent abstracts, and links to source records. The technology classification comes from analyzing patent titles and abstracts against a taxonomy of ~50 tech areas.

A single company can produce signals across both subtypes if their patent activity spans core product R&D and research-to-product transitions simultaneously.

📎

See real delivered data → Sample Files

Subtypes distinguish between patents in a company's core product area versus patents that signal a shift from research to product development:

Subtype EnumDescription
newCorePatentNew patents in the company's core product area — signals deepening investment in existing capabilities
productizationResearch-to-product shift detected — patents moving from theoretical to applied, signaling upcoming product launches

Example Signal

What a single entry looks like in a delivered signal file:

{
  "signal_id": "f1a3c5e7-9b2d-4f6a-8c0e-2d4f6a8b1c3e",
  "batch_id": "2026-03-15-00-00-00",
  "signal_type": "patent-filing",
  "signal_subtype": "newCorePatent",
  "detected_at": "2026-03-14T08:15:42.337Z",
  "association": "company",
  "company": {
    "name": "CrowdStrike Holdings",
    "domain": "crowdstrike.com",              // match on domain
    "linkedin_url": "linkedin.com/company/crowdstrike",  // or match on LinkedIn URL
    "industries": ["Cybersecurity", "Software Development"],
    "employee_count_low": 5001,
    "employee_count_high": 10000
  },
  "contact": [],
  "data": {
    "detail": "CrowdStrike filed 14 patents in a 60-day window focused on autonomous threat containment — AI systems that isolate compromised endpoints and remediate without human intervention. This cluster signals a shift from detection-focused...",
    "relevance": 0.94,                        // 0.0-1.0; higher = more actionable for outreach
    "confidence": "high",                     // how certain this signal is accurate
    "source_url": "https://assignment.uspto.gov/patent/index.html#/patent/search/result?searchInput=CrowdStrike",
    "tech_areas": [
      "Autonomous Threat Response",
      "Endpoint Isolation",
      "AI/ML For Security",
      "Behavioral Analysis",
      "Zero Trust Architecture"
    ],
    "uspto_assignee": "CrowdStrike, Inc.",
    "patent_count_recent": 14,
    "newest_patent_date": "2026-03-10T00:00:00Z"
  }
}

Field Reference

Standard envelope and entity fields are shared across all signals — see Schema and Resolution. The fields below are specific to this signal:

Signal-Specific Fields

The data object contains everything unique to this signal type — the intelligence extracted from patent filings.

FieldTypeDescription
detailstringMulti-sentence analysis of the patent cluster and its commercial significance. Explains what technology areas the company is investing in, what the patents suggest about product direction, and why it creates an outreach opportunity. Typically 3–5 sentences synthesizing multiple patent abstracts
relevancefloat (0.0–1.0)How actionable this signal is for outreach. Higher values reflect larger patent clusters in commercially relevant tech areas. Lower values indicate smaller clusters or purely research-oriented filings. Useful for prioritization and filtering
confidencestringHow certain the system is that the patent cluster was correctly attributed and classified. high = clear assignee match and consistent tech area, medium = probable but assignee name required fuzzy matching, low = inferred with limited data. Useful for production filtering
source_urlstring (URL)Link to the USPTO patent assignment search page for this company. Useful for manual verification or displaying to end users
tech_areasarray[string]Technology areas covered by the patent cluster, classified from patent titles and abstracts against a taxonomy of ~50 categories (e.g., "AI/ML For Security", "Edge Computing", "Natural Language Processing"). Useful for routing signals to relevant sales teams or building tech-stack-based targeting
uspto_assigneestringPatent assignee name as recorded in USPTO. May differ from company trade name (e.g., "Walmart Apollo, LLC" vs. "Walmart"). Useful for cross-referencing with patent databases
patent_count_recentintegerNumber of patents filed in the rolling detection window. Higher counts indicate more concentrated R&D investment. Useful for sorting signals by magnitude
newest_patent_datestring (ISO 8601)Date of the most recent patent in the cluster. Useful for recency filtering and understanding how current the R&D activity is

Timing & Delivery

  • detected_at is when we processed the patent batch. Use newest_patent_date for when the most recent patent was actually filed.
  • One signal per subtype per company per rolling window. A company can have both newCorePatent and productization signals active simultaneously, but each subtype fires only once per detection cycle.
  • Each delivery arrives in a timestamped folder. Treat all signals in a new folder as recent — no need to diff against prior deliveries.

Coverage

  • Refresh: Monthly
  • Coverage: 2,000,000+ companies (USPTO only; international patent offices coming soon)
  • Best for: Enterprise sales to R&D-heavy companies, technology vendors, identifying companies investing in areas adjacent to your product

Contact Sales →

Updated 9 days ago