Patent Signals (Company)
Company-level US patent activity: grants, portfolio velocity, outside IP counsel, and PTAB challenges — enriched and delivered weekly.
Overview
Company-level US patent signals, refreshed weekly from the USPTO grant bulk file (PTGRXML,
published Tuesdays), PTAB proceedings, and the USPTO assignment feed. Each record is enriched
against Autobound's company truth data and delivered under signal_type: patents,
association: company. Use it to spot R&D investment, portfolio momentum, IP disputes, and
which outside counsel a company retains.
For the inventor (person) view of the same grants, see Patent Signals (Inventor).
Available Subtypes
Subtype (signal_subtype) | Description |
|---|---|
patentGrant | A newly granted US patent assigned to the company |
portfolioVelocity | Weekly snapshot of the company's patent output and trend |
patentChallenge | A PTAB proceeding (IPR/PGR) where the company is the petitioner |
usesLawFirm | The company's outside IP counsel, inferred from prosecution records |
Schema
Every record has top-level envelope fields, a company object, and a subtype-specific data
object.
Top-Level Fields
| Field | Type | Description |
|---|---|---|
signal_id | string | Stable ID (upsert on this) |
batch_id | string | Delivery batch |
signal_type | string | Always patents |
signal_subtype | string | One of the subtypes above |
detected_at | string (ISO 8601) | When the signal was produced |
association | string | company |
Company Object
Enriched from Autobound's company truth data. Unknown fields are null — never guessed.
| Field | Type | Description |
|---|---|---|
name | string | Assignee company name |
domain | string | Company domain |
linkedin_url | string | null | Company LinkedIn URL |
industries | string[] | Industry labels |
employee_count_low | integer | null | Employee band low |
employee_count_high | integer | null | Employee band high |
description | string | null | Company description |
Data Object
Prose fields on the LLM-backed subtypes (patentGrant, patentChallenge): summary,
detail, invention_explanation, technologies_mentioned (string[]), tags (string[]).
Every subtype carries source_url, uspto_assignee, event_at, event_kind.
patentGrant — patent_number (US11234567), patent_title, patent_abstract,
grant_date, cpc_primary, cpc (≤5), tech_area, inventors[]
({ name, linkedin_url, is_primary, status }). event_kind = grant_date.
portfolioVelocity — window_asof_date, grants_30d, grants_60d, grants_90d,
grants_180d, grants_365d, grant_velocity_ratio, grant_direction
(accelerating/steady/decelerating), challenges_made and challenges_received
(each { "90d", "365d" }), unique_inventors, primary_cpc_subclass, cpc_section_mix,
top_tech_areas[]. event_kind = velocity_snapshot.
patentChallenge — company is the petitioner. trial_number, trial_type,
trial_status, challenge_role, patent_owner_name, patent_owner_domain,
petitioner_counsel_name, challenged_patent_number, challenged_patent_title,
challenged_patent_inventors[], petition_filing_date, institution_decision_date,
latest_decision_date, termination_date, technology_center_art_unit, why_this_fight,
challenge_context, competitors_mentioned[] (the patent owner). event_kind is the stage:
petition_filed/instituted/decided/terminated.
usesLawFirm — law_firm_name, law_firm_domain, relationship_basis (prosecution),
attorney_name, representative_matter, matter_date. event_kind = representation.
Example (patentGrant)
patentGrant){
"signal_id": "b1f0…",
"batch_id": "patents-2026-09-06",
"signal_type": "patents",
"signal_subtype": "patentGrant",
"detected_at": "2026-09-06T07:03:11Z",
"association": "company",
"company": {
"name": "Acme Robotics",
"domain": "acmerobotics.com",
"linkedin_url": "https://www.linkedin.com/company/acme-robotics",
"industries": ["Robotics"],
"employee_count_low": null,
"employee_count_high": null,
"description": "Autonomous warehouse robotics."
},
"data": {
"summary": "Acme Robotics was granted a US patent on adaptive gripper control.",
"patent_number": "US11987654",
"patent_title": "Adaptive gripper control for variable-geometry payloads",
"grant_date": "2026-09-01",
"cpc_primary": { "code": "B25J9/16", "label": "Program-controlled manipulators" },
"tech_area": "Robotics & Automation",
"technologies_mentioned": ["force feedback", "grasp planning"],
"inventors": [
{ "name": "Dana Ng", "linkedin_url": "https://www.linkedin.com/in/dana-ng", "is_primary": true, "status": "resolved" }
],
"event_at": "2026-09-01",
"event_kind": "grant_date",
"source_url": "https://patents.google.com/patent/US11987654"
}
}
No scoring fieldsPatent signals do not carry
relevance,confidence, orsentiment. Every delivered
field is factual (from USPTO/PTAB) or LLM-written prose — nothing is scored.
Migrating from patents / Patent Filings
patents / Patent FilingsThis replaces the legacy single Patent Filings signal (type patents,
subtypes newCorePatent/productization). It is split into patents-company and
patents-contact, and the payload is the enriched schema above. The old slugs — patents,
patent-filing, patent — still resolve as aliases to patents-company so existing
queries keep working, but they are deprecated and targeted for removal ~2026-10-09. Move
filters to patents-company (and patents-contact for inventors) before then.
Coverage
- Bucket:
autobound-patents-company(mirrored to your S3 delivery bucket where configured) - Refresh: weekly (USPTO grants publish Tuesdays)
- Association: company
- Subtypes: 4 (
patentGrant,portfolioVelocity,patentChallenge,usesLawFirm)
Updated 6 minutes ago

