Podcasts

Podcast appearances, in-episode business claims with verbatim quotes, and sponsorship ad reads — company and contact associations, refreshed weekly.

Podcast signals turn transcribed B2B podcast episodes into three kinds of structured rows: appearances (a tracked contact or company shows up as guest/host, with talking points and a company snapshot), claims (specific business events or statements — funding, tech adoption, pain points, executive opinions — each backed by verbatim transcript quotes), and sponsorships (verbatim ad reads, mapped to the sponsoring company).

Every row uses the standard signal envelope (signal_id, signal_type, signal_subtype, association, detected_at, contact, company, data) — see Schema. Match on company.domain, contact.email, or linkedin_url per Resolution.

📎

See real delivered dataSample Files

Two Signal Types

signal_typeassociationWhat it carries
podcast-contactcontactA person's podcast appearances and the claims they made on-air
podcast-companycompanyCompany appearances, claims about the company, and sponsorship ad reads

Subtype Families

Three payload shapes, distinguished by signal_subtype:

FamilySubtypesWhat it is
AppearancepodcastAppearance (contact), companyPodcastAppearance (company)One row per entity per episode: who appeared, what they talked about, snapshot of their company
Claim65+ subtypes — aiInvestment, techAdoption, painPointDisclosed, strategicInitiative, marketExpansion, careerHistory, fundingEvent, executiveOpinion, industryPrediction, ...One row per distinct business claim made in an episode, with verbatim quotes
SponsorshippodcastSponsorshipOne row per sponsor ad read, with the verbatim offer text

The claim subtype vocabulary is shared with SEC filing and earnings-call signals, so a signal_subtype filter like aiInvestment works identically across sources. Full list → Signal Catalog.


Payload Reference: Appearance

Full signal, podcast-contact / podcastAppearance:

{
  "signal_id": "dca86039-9c23-54d3-b9c2-00eacdb38baf",
  "signal_type": "podcast-contact",
  "signal_subtype": "podcastAppearance",
  "association": "contact",
  "detected_at": "2026-08-24T10:25:28Z",
  "contact": {
    "full_name": "Arundhati Bhattacharya",
    "first_name": "Arundhati",
    "last_name": "Bhattacharya",
    "job_title": "Chairperson & CEO, Salesforce India",
    "email": "[email protected]",
    "linkedin_url": "https://in.linkedin.com/in/arundhati-bhattacharya-salesforce"
  },
  "company": {
    "name": "Salesforce",
    "domain": "salesforce.com",
    "linkedin_url": "linkedin.com/company/salesforce"
  },
  "data": {
    "podcast_name": "CII Podcasts",
    "episode_title": "India@100: Powering the Future through Digital Acceleration ft Arundhati Bhattacharya",
    "episode_url": "https://shows.acast.com/ciipodcasts/episodes/india100-powering-the-future-through-digital-acceleration-ft",
    "episode_id": "55821260383",
    "published_at": "2026-06-05",
    "headline": "Arundhati Bhattacharya (President & CEO, Salesforce, South Asia @ Salesforce) on CII Podcasts",
    "host": "Rajan Navani",
    "occasion": "Discussing the CII's 'India@100' initiative",
    "topics": [
      "India's economic future",
      "Digital transformation",
      "AI adoption",
      "Cloud computing",
      "Public-private partnerships",
      "Technology policy"
    ],
    "talking_points": [
      {
        "point": "India must become a creator of technology, not just a consumer.",
        "quote": "I also want to see India not only as a consumer of technology, but actually as a creator of technology. India has long been a consumer and yet, you know, we are one of the most digitally empowered societies."
      },
      {
        "point": "Board-level 'fear of missing out' is driving AI investment.",
        "quote": "once you realize that it is something that's here to stay, then there is a huge fear of FOMO that builds up... And when that happens, they then actively start looking at how they should also be an early beginner rather than a late one."
      }
    ],
    "company_snapshot": {
      "what_they_do": null,
      "org_facts": [],
      "technologies_used": [],
      "customers_mentioned": [],
      "competitors_mentioned": []
    },
    "transcript_source": "transcript",
    "recorded_at_estimate": "2024-06",
    "date_confidence": "inferred",
    "is_ad": false
  }
}

(Real production signal. Note the company_snapshot here is empty — it's populated when the episode substantively discusses the company itself; interview-style appearances about broader topics often leave it sparse. recorded_at_estimate shows a 2024 date because this episode is a republished recording — exactly the drift the field exists to surface.)

FieldTypeDescription
podcast_namestringShow name
episode_titlestringEpisode title
episode_urlstring, nullableLink to the episode — may be null when no canonical URL was resolvable
episode_idstring, nullableStable episode identifier — join/dedupe key across rows from the same episode. Occasionally null on sponsorship rows
published_atstring (date)Episode publication date
headlinestringOne-line summary of the appearance
hoststring, optionalShow host
occasionstring, optionalWhy the guest was on (launch, book, event, series)
co_guestsarray, optionalOther guests on the episode
topicsarray[string]Topics discussed
talking_pointsarray[object]What the entity said, as {point, quote} pairs — point is the paraphrase, quote is verbatim from the transcript
company_snapshotobjectCompany picture as described in the episode: what_they_do, org_facts[], technologies_used[], customers_mentioned[], competitors_mentioned[] — arrays are empty when the episode doesn't cover them
transcript_sourcestringAlways "transcript" — extraction ran on the full episode transcript, not show notes
recorded_at_estimatestring (date or year-month)Estimated recording date, e.g. "2026-07" — episodes are often published weeks (or reruns: years) after recording
date_confidencestring"exact" or "inferred" — confidence in the recording-date estimate
is_adbooleantrue only when the row comes from a house ad (the show promoting its own product); regular appearances are false

Payload Reference: Claims

Same envelope; data payload for a real claim row (subtype aiInvestment — John Deere's CTO on Masters of Scale):

{
  "headline": "John Deere · aiInvestment · in_progress (2026-07)",
  "quotes": [
    "The generative AI and I would argue transformer networks in general are interesting to us for a whole variety of reasons.",
    "But generative models have given us the ability to sort of reject the noise in the data and focus on the signal, which is and to be able to do that at faster clock speeds than we've traditionally been able to do it.",
    "We're also interested, though, in them for edge use cases. I talked to you about the autonomous use case as an example."
  ],
  "evidence": "The generative AI and I would argue transformer networks in general are interesting to us for a whole variety of reasons.",
  "speaker": "Jahmy Hindman",
  "speaker_relation": "own_company",
  "timing": "in_progress",
  "event_date": "2026-07",
  "podcast_name": "Masters of Scale",
  "episode_title": "Pioneers of AI: John Deere's AI vision for future farms",
  "episode_url": null,
  "episode_id": "57408329498",
  "published_at": "2026-07-11",
  "date_confidence": "inferred",
  "recorded_at_estimate": "2026-07",
  "transcript_source": "transcript",
  "is_ad": false
}
FieldTypeDescription
quotesarray[string]Verbatim transcript quotes supporting the claim
evidencestringConvenience copy of the first quote
speakerstringWho said it
speaker_relationstringown_company (speaker talking about their own company) or third_party (someone else talking about the associated entity)
timingstringWhen the claim applies — e.g. in_progress, ongoing_state, imminent
event_datestring (date or year-month), optionalPresent when a concrete date is stated or inferable for the event
headlinestringOne-line claim summary in the form Company · subtype · timing (date)
Episode fieldspodcast_name, episode_title, episode_url, episode_id, published_at — same as appearance rows

A single episode typically yields one appearance row plus several claim rows; join them on episode_id. Filter speaker_relation = "own_company" for first-party disclosures.

Payload Reference: Sponsorship

data payload for podcast-company / podcastSponsorship:

{
  "headline": "Bill sponsors Accounting Today Podcast",
  "offer_texts": [
    "Trusted by 98 of the top 100 U.S. firms, Bill provides better visibility into payment status and cash flow. Use AI to simplify your tech stack and move your firm to a new level of productivity. Take a demo at Bill.com slash accounting today, and you'll get a $100 Amazon gift card. Terms apply."
  ],
  "is_ad": false,
  "podcast_name": "Accounting Today Podcast",
  "episode_title": "Predicting the next 15 years in accounting",
  "episode_url": "https://www.accountingtoday.com/podcast/predicting-the-next-15-years-in-accounting",
  "episode_id": null,
  "published_at": "2026-06-15",
  "date_confidence": "inferred",
  "recorded_at_estimate": "2026-06",
  "transcript_source": "transcript"
}
FieldTypeDescription
offer_textsarray[string]Verbatim ad reads — includes promo codes, offer terms, landing pages
headlinestringOne-line summary: who is sponsoring what
is_adbooleantrue only for house ads (the show promoting its own product); third-party sponsor reads are false
Episode fieldsSame as above — episode_id is often null on sponsorship rows

Useful for ad-spend and share-of-voice analysis: which companies are buying reads on which shows, and what they're offering.


Freshness & Scale

  • Refresh: Weekly (Sunday batches). Every batch is reviewed before publish.
  • Current volume: ~59K signals — 49,947 podcast-contact + 8,980 podcast-company — across ~27,700 companies, extracted from ~39K transcribed episodes.
  • detected_at is when the batch published; use published_at and recorded_at_estimate in data for episode timing.

Transcript Keyword Search

Both podcast types are content-searchable via content_filters on POST /v1/contacts/search and POST /v1/companies/search. The keywords clause partial-matches (case-insensitive) across headlines, evidence, quotes, talking-point points and quotes, sponsor offer reads, and topics.

curl -X POST https://signals.autobound.ai/v1/contacts/search \
  -H "X-API-KEY: $AUTOBOUND_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "signal_types": ["podcast-contact"],
    "detected_after": "2026-08-01",
    "content_filters": {
      "keywords": { "contains": ["digital transformation", "AI adoption"] }
    }
  }'
📘

keywords only. For podcast types, keywords is the only content_filters field that matches — tech_mentioned, competitors_mentioned, pain_points, and initiatives run against extracted fields that exist only on social signal types and will never match a podcast row. Mixing podcast types with any non-content-searchable signal type in the same request still returns a 400.


Delivery

Same channels as every other signal type: query via the Search and Enrich APIs, or receive flat files in timestamped GCS folders (JSONL/Parquet) — see Delivery and Sample Files.


Contact Sales →


Did this page help you?