Skip to main content
Product update · Fall 2026

Every podcast is a
buying signal waiting to be heard.

A VP announces nine open roles. A CEO names the POS system they just switched to. Ford's CEO says the platform cost has to come out. Out loud, for 45 minutes, and it evaporates when the episode ends. We turn 12,000 episodes a day, and conference stages too, into typed, timed, attributed signal. Here's how it works, the best of what it caught in the last few weeks, and what's coming next.

Signal pipeline
The Dave Gerhardt Show (Exit Five)
Marketing to B2B buyers who aren't on LinkedIn · Aug 24, 2026

transcript Now we have 17 full-time employees, three to four contractors, have nine new roles in the second half of the year for marketing, just the marketing team alone. So we're growing really, really fast.

The signal layer, by the numbers

700+
Signals across 35+ sources
250M+
Contacts
50M+
Companies
<200ms
Response time (p95)
Changelog · Jun 4 to Sep 1

What shipped this quarter

A lot. Four headlines first, then the full log: 30 dated entries across sources, API, delivery, and quality. Customer-facing changes only, dated by the day they reached production. The signal catalog has the complete list of types.

Shipped Jun 17

Podcast intelligence

Two signal types, podcast-contact and podcast-company, mined from what executives say on air. We poll 30,000+ B2B feeds daily and process about 12,000 episodes a day into ~95 subtypes, each with a verbatim quote and the speaker resolved to a real contact. Transcript keyword search landed on Aug 25 via content_filters.keywords, and a podcastAppearance subtype for guest appearances on Sep 1.

podcast signals delivered
70,000+
podcast signals delivered
episodes a day
~12,000
episodes a day
carry the speaker's LinkedIn URL
100%
carry the speaker's LinkedIn URL
Shipped Jun 4

Conference speaker signals

conference-speaker reads agendas from roughly 5,400 conferences a year and pulls the session title, track, event, location, and recording when one exists, with the speaker resolved to a contact and company. An 18-month backfill delivered 13,500+ speaker signals from 4,700+ distinct speakers, and every one carries the conference URL and a valid date.

conferences a year
~5,400
conferences a year
speaker signals backfilled
13,500+
speaker signals backfilled
carry conference URL + valid date
100%
carry conference URL + valid date
Shipped Jun 16

Monitoring pools via API

POST /v1/monitoring/add replaces the old CSV hand-off. Send 1 to 100 records per call and they are watched across every signal type we run. 10 credits per net-new entry, and duplicates are free, so re-sending your book costs nothing.

records per call
1-100
records per call
credits per net-new entry
10
credits per net-new entry
credits for duplicates
0
credits for duplicates
Shipped Jun 16

Company-level buyer intent

POST /v1/buyer-intent/companies with keyset pagination and firmographic filters, plus a free GET /v1/buyer-intent/companies/{domain}/timeline and async exports for both contacts and companies. Contact intent moved to Postgres on Jun 15 and now serves sub-100ms lookups.

contact intent lookups
<100ms
contact intent lookups
company timeline endpoint
Free
company timeline endpoint
exports, contacts + companies
Async
exports, contacts + companies

The full log

30 entries

Sources & data

  • Jun 8

    News delivered daily instead of weekly.post →

  • Jul 10

    News v4 schema: 40 subtypes and 16 new fields including sentiment, confidence, and a related-company block. Every existing field preserved.post →

  • Jul 21

    GitHub signals now produced in-house from the GitHub API.docs

  • Jul-Aug

    Hiring Velocity and Hiring Trends rebuilt on a new in-house pipeline. A single weekly drop now carries 220,000+ velocity and 370,000+ trends signals.docs

  • Aug 3

    Industries normalized to one canonical taxonomy across every signal, replacing hundreds of inconsistent values.changelog

  • Aug 18

    SEC 6-K now reads EX-99 exhibits, so foreign filers that previously yielded zero signals now do.docs

  • Sep 1

    SEC Form D domain resolution up from 21% to 71%.docs

Quality & speed

  • Jun-Aug

    The search timeout class is closed. Keyword searches that took 60s+ now return in milliseconds, behind a hard 9.5s deadline with an honest 408 and suggested fixes.

  • Jul 6

    Reddit buying intent gated to B2B. 97% of prior matches were consumer.docs

  • Jul 6

    Glassdoor: domain-verified, no rating-only signals, confidence tied to review count.docs

  • Jul 21

    Job-change pipeline: three systemic bugs fixed that had affected the majority of job-change records.

  • Aug 5

    Earnings extraction v8: filler regions 34% → 0%, duplicate quotes 4% → 0%, and vendors and regions mentioned are now captured.docs

  • Aug 16

    Proactive freshness monitoring on every delivered signal type.

  • Aug 21

    Every one of 34 signal types enumerates its subtypes in under a second.docs

  • Aug 27

    LinkedIn URL mismatches (5-8% of records) fixed with match-or-abstain resolution.

API

  • Jun 4

    content_filters on company and contact search: keywords, tech mentioned, competitors mentioned, pain points, and initiatives, each with contains / not_contains, across LinkedIn, Twitter/X, and YouTube signals.post →

  • Jun 11

    Machine-readable 400s. Invalid signal types return the invalid and valid lists, and URLs normalize to domains.

  • Jul 7

    Contacts carry first_name, last_name, full_name, job_title, city, state, and country.docs

  • Jul 10

    Email verification: POST /v1/emails/verify at ~385ms, and bulk up to 1,000 per call. 1 credit per email, unknown results free.post →

  • Jul 14

    Default search window widened from 30 to 50 days.changelog

  • Jul 29

    Contact search filters on canonical seniority, department, and job title.changelog

  • Aug 20

    signal_subtypes filter on search: up to 25 values, OR across types, with did-you-mean on unknowns. The subtype vocabulary grew from 730 to 962 values across 34 types.docs

  • Aug 21

    GET /v1/filters: free canonical picklists for industry, country, seniority, department, and revenue. Bad values now hard-400 with suggestions instead of silently returning empty.docs

  • Aug 27

    signal_subtypes on enrich endpoints too. Filtered-out signals are not billed.changelog

  • Jun-Aug

    MCP server: buyerIntentCompanies, buyerIntentCompanyTimeline, verifyEmail, verifyEmailBulk, getFilters, and addToMonitoringPool added. Sandbox tools 8 → 18. OpenAPI spec 2.1.0 → 2.4.0.docs

Delivery & billing

  • Jun 9

    Delivery folders UTC-timestamped at production time, with a dedup ledger and straggler handling so no signal is missed across drops.changelog

  • Jul 2

    Auto top-up for all workspaces: saved card, threshold, receipts, optional monthly cap.post →

  • Jul 6

    Enterprise credit packs: $10K, $15K, $20K, $40K.changelog

  • Jul 16

    Push delivery to customer-owned S3 buckets, alongside GCS push and flat file.docs

  • Aug 23

    Earn credits: G2 review, referral, first API call, webhook connected.

Deep dive

The most robust signal we've built

Podcasts are the richest unstructured buying-intent source on the internet, and almost nobody mines them, because doing it well is genuinely hard. Here's the pipeline, end to end.

  1. 01

    Discover

    Poll 30,000+ B2B podcast feeds daily, ranked by a value score so the highest-signal shows get processed first.

  2. 02

    Transcribe

    Speech-to-text on every episode, corrected against episode metadata so names and companies aren't mangled.

  3. 03

    Extract

    An LLM reads the full transcript and pulls 8-15 discrete signals from a single 30-minute episode.

  4. 04

    Structure

    Each signal is tagged against a taxonomy of ~95 subtypes, with a verbatim quote, a confidence score, and a relevance score.

  5. 05

    Resolve

    Every named person and company is matched against a 950M-record identity graph, resolving to a real LinkedIn, email, title, and domain.

  6. 06

    Validate & ship

    The whole day's batch runs 9 hard gates: zero taxonomy violations, sub-2% duplicates, minimum domain fill. A failed gate ships nothing.

The result: signals carry full provenance. Podcast name, episode, publish date, source URL, whether the date is exact or inferred, and whether the speaker was talking about their own company or someone else's. We'd rather delay than deliver garbage.

From the transcript · podcast-contact

Typed. Timed. Attributed.

We reviewed 1,100+ podcast signals from the last few weeks for this post and picked six. Every one is first-person: the speaker talking about their own company. That is the podcast-contact type; companies mentioned by other people land in podcast-company. Each card is the signal as delivered: the subtype, the verbatim quote, the episode it came from, the publish date, the timing, and the executive resolved to a company. Transcript keyword search is live too, via content_filters.keywords changelog ↗.

podcast signals delivered
70,000+
podcast signals delivered
episodes across 4,700+ podcasts
17,500+
episodes across 4,700+ podcasts
speakers
15,000+
speakers
verbatim quotes
182,000+
verbatim quotes
carry the speaker's LinkedIn URL
100%
carry the speaker's LinkedIn URL
company domain fill, latest delivery
100%
company domain fill, latest delivery
carry a verbatim transcript quote
99.7%
carry a verbatim transcript quote
duplicate signal IDs
0
duplicate signal IDs
legacyModernization
Legacy modernization signal
transcript
We're in the pilot stages right now, which is exactly where we should be. We're learning, measuring, iterating. We're making sure the model's right before we scale rapidly. This year, we're going to complete our first 200 transformations.

At Your Convenience · 7-Eleven exec shares retailer's expansion plans · Aug 25, 2026 · in progress

Terri Micklin
SVP, Store Development & Facilities, 7-Eleven
✓ resolved7-eleven.com
10,000+ employees · Retail

Why it matters. 200 store remodels this year, then rapid scale. Fixtures, construction, POS, signage, and facilities vendors have a dated trigger and a named executive.

automationInvestment
Automation investment signal
transcript
We are hoping to expand the use of agents to try and fill the gap of all of the new work that's coming in and all of the extended modules that we're looking at without changing the headcount.

The Alldus Podcast - AI in Action · Building a Scalable ServiceNow Platform for Healthcare Innovation with Wellstar's Dianne Kokotoff · Aug 24, 2026 · in progress

Dianne Kokotoff
Executive Director, Enterprise Solutions and Automation, Wellstar Health System
✓ resolvedwellstar.org
10,000+ employees · Hospitals and Health Care

Why it matters. A 25,000-person health system buying AI agents and new platform modules under a flat-headcount mandate. That is a budget with a shape.

digitalTransformation
Digital transformation signal
transcript
So now we're thinking about how do you not only just modernize our capabilities, move to the cloud, use SaaS platforms, but really simplify and challenge every process as part of that.

Technovation with Peter High · Simplify First, Scale AI Next with Vulcan Materials CIO Krzysztof Soltan · Aug 24, 2026 · in progress

Krzysztof Soltan
Chief Information Officer, Vulcan Materials Company
✓ resolvedvulcanmaterials.com
10,000+ employees · Construction materials (NYSE: VMC)

Why it matters. A Fortune 500 CIO stating the roadmap sequence out loud: commercial first, then finance, HR, and supply chain. You know which door opens next.

techAdoption
Tech adoption signal
transcript
We're huge Gong users, and I'm a big fan of Gong. No special plug for them.

That's What I Call Marketing · Canva B2B CMO on B2B Marketing, Enterprise Growth and Building Demand · Aug 24, 2026 · ongoing

Meghan Gendelman
CMO, B2B, Canva
✓ resolvedcanva.com
5,001-10,000 employees · Software

Why it matters. A named revenue-intelligence vendor at a top-tier brand, from the executive who owns the B2B motion. Displacement target for one side, expansion for the other.

painPointDisclosed
Pain point disclosed signal
transcript
That doesn't mean my whole team is closing the books faster yet. We're on the path there and I see the light, but I don't have the system in place yet to make that kind of end-to-end accounting and finance process faster than it was before.

CFO Thought Leader · 1209: Finding a Larger Voice in the Room | Ross Grainger, CFO, Nasuni · Aug 26, 2026 · ongoing

Ross Grainger
Chief Financial Officer, Nasuni
✓ resolvednasuni.com
1,001-5,000 employees · Hybrid cloud storage

Why it matters. A CFO saying he lacks the system for end-to-end close automation. Textbook opening for FP&A, close, and AI-finance vendors.

aiInvestment
AI investment signal
transcript
Right now, we're in the beginning stage of completely redoing our website and starting from scratch with an agentic AI native website with our own MCP that will allow us to connect directly to Claude and ChatGPT and get our booking links in front of the OTAs.

No Vacancy Live! · FNA 217: Belon Our Wildest Dreams · Aug 9, 2026 · in progress

Dina Belon
President, Staypineapple Hotels
✓ resolvedstaypineapple.com
501-1,000 employees · Hospitality

Why it matters. A hotel group president announcing a from-scratch web rebuild with MCP and agent integrations. An RFP-shaped signal for web, booking-engine, and AI vendors.

Two more from the same run
  • cybersecurityInvestmentCybersecurity investment

    Whenever we acquire a new company... we put our security wrapper around those companies.

    Nikhil Narvekar, CIO, Graphic Packaging
  • strategyShiftStrategy shift

    suddenly we can play offense again. We can increase marketing budgets to reach to a broader group of consumers.

    Bernard Osta, Executive Director, Vestiaire Collective
    Marketing in the Madness · Aug 25, 2026
New signal type · conference-speaker

From the stage

Who an account puts on stage tells you what it is prioritizing. We read conference agendas, roughly 5,400 conferences a year through a Sessionize fast path plus an agentic crawl of agenda pages, pull speaker, session, track, and event metadata, and resolve the speaker to a contact and company. An 18-month backfill delivered 13,500+ speaker signals from 4,700+ distinct speakers, and every one carries the conference URL and a valid date. Six from the backfill, in order.

conferences a year tracked
5,400
conferences a year tracked
speaker signals, 18-month backfill
13,500+
speaker signals, 18-month backfill
distinct speakers
4,700+
distinct speakers
carry the conference URL and a valid date
100%
carry the conference URL and a valid date
duplicate signal IDs
0
duplicate signal IDs
  1. 03
    Mar 2026
    Talk

    Evaluating Threats & Automating Defense: How Google is Advancing Code Security

    [un]prompted II · San Francisco, CA · Stage 1

    Our discussion will focus on advancing code security, provide a comprehensive overview of Google's AI security strategy, show how we evaluate emerging cyberattack capabilities and demonstrate how tools like CodeMender are helping build intrinsically safer software.

    with Four Flynn, VP Security and Privacy, Google DeepMind

    conferenceSpeaker
    Conference speaker signal
    Heather Adkins
    VP of Security Engineering, Google
    ✓ resolvedgoogle.com

    Why it matters. Two Google security VPs naming a specific internal tool. Roadmap disclosure for AppSec vendors, and a warm co-speaker pair.

  2. 10
    Mar 2026
    Keynote

    Innovating at Pace with Simpler Decision-Making

    CTO Craft Con: London · London, UK · ~500 attendees

    Matej Pfajfar, CTO of Monzo, the UK's largest digital bank, shares insights from a career spanning leadership roles at Monzo, Google, and high-growth startups. Drawing on his experience enabling machine learning at scale, leading Android platform innovations, and building resilient systems for millions of users.

    conferenceSpeaker
    Conference speaker signal
    Matej Pfajfar
    CTO, Monzo
    ✓ resolvedmonzo.com

    Why it matters. The CTO of a 10M-customer bank on ML at scale and resilience. Right persona, right themes, fresh reason to reach out.

  3. 28
    Apr 2026
    Keynote

    Keynote Interview: Vercel's Guillermo Rauch on the Agent Era

    Azure Cosmos DB Conf 2026 · Virtual · 500-2,000 attendees

    Guillermo Rauch, CEO of Vercel and creator of Next.js, joins Kirill Gavrylyuk on building a developer cloud for the AI agent era.

    conferenceSpeaker
    Conference speaker signal
    Guillermo Rauch
    CEO, Vercel
    ✓ resolvedvercel.com

    Why it matters. A unicorn CEO keynoting a Microsoft data-platform event. A deepening partnership, visible to every partner and competitor in that ecosystem.

  4. 04
    May 2026
    Keynote

    Power the agentic enterprise

    IBM Think 2026 · Boston, MA · ~4,600 attendees

    with Rob Thomas, SVP Software & CCO, IBM · Detlev Klage, Vice Chairman, Finanz Informatik · Dinesh Nirmal, SVP Software, IBM

    conferenceSpeaker
    Conference speaker signal
    Suresh Viswanathan
    Group Chief Operating Officer, Nationwide Building Society
    ✓ resolvednationwide.co.uk

    Why it matters. A UK bank's COO on IBM's main stage about agentic AI. The account is mid-deployment, and adjacent vendors have a dated, sourced trigger.

  5. 11
    May 2026
    Talk

    Making changes at Scale: large scale roll out of OpenRewrite and Moderne one year later

    Code Remix Summit Miami · Miami, FL · 500-1,000 attendees

    Morgan Stanley has been leveraging Moderne for close to a year now as part of its suite of modernization tools to automate large-scale refactoring and technical debt remediation. Join us to learn more about our approach, our successes, and our challenges to date, and what we have planned next to continue this transformation.

    with Khalid Elsawaf, Executive Director, Morgan Stanley

    conferenceSpeaker
    Conference speaker signal
    Dov Katz
    Managing Director, Distinguished Engineer, Morgan Stanley
    ✓ resolvedmorganstanley.com

    Why it matters. A named enterprise confirming a year-old vendor rollout and "what we have planned next". Live budget, two named buyers.

  6. 11
    Aug 2026
    Talk · Focus Track

    Move Fast and Don't Break Things: A New Model for AI Governance

    AI Risk Summit 2026 · Half Moon Bay, CA · 300-500 attendees

    This session explains how enterprises can move faster with AI without sacrificing control by rebuilding governance and authorization models to match the speed of AI agents. It focuses on replacing static policy with controls that keep pace with every AI decision and provide a complete record of actions.

    conferenceSpeaker
    Conference speaker signal
    Art Gilliland
    CEO, Delinea
    ✓ resolveddelinea.com

    Why it matters. A security CEO publicly repositioning around agent authorization. Every identity partner and competitor should know.

Same shape as every other signal: a subtype, a resolved contact and company, the source event, and the date. Different medium, identical output. Association is at the contact level, so the person on stage is the person you reach.

Roadmap

What's coming

Where we're taking the Signal API next. Monitoring pools are the headline, and everything underneath makes them sharper and easier to build on.

Monitoring pools that push

This is the big one. A monitoring pool is the exact set of companies and contacts you care about, your whole book, a target list, one strategic account. Today you build a pool and query it. Next, it becomes a living feed: pools that grow themselves, watch every source we run, and push you the moment a watched entity moves. It turns the Signal API from something you check into something that reaches out to you.

  1. Pick who to watch
  2. Pool grows + monitors
  3. Signal fires
  4. Pushed to your CRM
01

Pools that grow themselves

Add 10,000 accounts once and the pool keeps widening on its own, resolving net-new contacts at each company and folding in newly matched entities as they surface. Coverage compounds instead of decaying the week after you upload a list.

02

Every source, one pool

One pool watches news, podcasts, conference agendas, Reddit, hiring, earnings, SEC filings, and web activity at the same time. We are scaling pool coverage across all 35+ sources so a single watchlist sees everything, not one channel at a time.

03

Push, not poll

When a watched entity fires a new signal, it lands on your webhook or writes into your CRM in near real time, already resolved and attributed. Stop cron-polling for changes and start getting notified the instant they happen.

Underneath the pools

One universal schema

Every source, from a 10-K to a podcast to a Reddit thread, resolving to the same shape: same fields, same provenance block, same date logic, same confidence semantics. Write your integration once and every signal type, current and future, drops straight in with no per-source special-casing.

Sharper entity resolution

A shared resolution service and custom ML models sit behind every pipeline, so each signal attaches to the right person at the right company instead of a near-match. Pass a LinkedIn URL or a domain and get back one clean, confident entity with a score you can filter on.

Verified before it ships

Every delivery gets sampled against ground truth before release, and every pipeline keeps its raw source data so any signal can be traced back and re-derived. Quality gates you can see, not just trust.

More sources

Patent filings and a rebuilt Reddit pipeline are landing first, with richer website-intelligence synthesis and SEC Form D private-funding signals close behind, so you catch raises and intent before they hit the press.

Every signal we ship should look the same on the wire, whether it came from a 10-K, a podcast, or a Reddit thread. And it should land on the right person at the right company, every time. Get those two things right and everything downstream, monitoring pools, pushes, your own models, just works. That's why I'm more excited about schema and resolution than about any single new source.
Daniel Wiener, CEO, Autobound

Want to see it in your account?

700+ signals across 35+ sources, full provenance, and a podcast pipeline pulling typed, reachable buying signals out of 12,000 episodes a day.