What "verified" means here.
The largest structured archive of what executives say when they are not reading from a script. This page explains how a two-hour podcast becomes a record you can cite, and who is responsible for it.
How an appearance becomes a record
Part of this work is automated and we say so. A human review sits in the middle of every record before it reaches a customer.
- Continuous sourcingWe scan YouTube and the web up to five times a day for new appearances by the 34,000+ executives we track, plus channel-level onboarding of podcasts and conference series.
- Model pre-vettingA first pass checks relevance and that the speaker named in the source is who the video shows. Obvious mismatches never enter the pipeline.
- Secure ingest with provenanceThe appearance and its metadata land in PostgreSQL with a link back to the original source and the exact timestamp of every turn.
- Human quality control and compliance reviewReviewers confirm the executive is the one speaking, screen for personal information and material nonpublic information, and reject dubbed audio, clips of clips, and secondhand commentary.
- StructuringSpeaker attribution per turn, entity IDs on the executive's own turns, cleaned sentences, an appearance date resolved from the content rather than the upload date, topic and event-type tags, and extracted quotes.
- Deduplication and releaseReposts of the same conversation collapse to the authoritative source. Only then does the record reach the app and the API.
What you get on every record
- The executive's own turns, isolated from the host, with their entity ID.
- A timestamp on every turn, linked to the second in the source.
- An appearance date that reflects when it was said, not when it was uploaded.
- Company, ticker, and current title. Event type: podcast, conference, TV, earnings.
- Extracted quotes with financial-policy and controversy flags.
Why this exists
Around 2020 the people running the largest companies in the world started sitting for two and three hour conversations on podcasts, roundtables, and industry keynotes. They walk through roadmaps and give real opinions on export controls, immigration, and supply chains. None of it is in a 10-K, and none of it is on a terminal. We verified roughly 55,000 executive appearances in 2025, three times the 2024 count and more than every year from 2006 through 2019 combined. CEOInterviews.AI captures that channel and structures it so you can search ten years of an executive's own words in seconds.
The team
CEOInterviews.AI is a US company with offices in Los Angeles, San Francisco, and Hong Kong. More than twenty labelers and annotators work full time on data quality, and we work directly with the funds and universities that depend on it. Security reviews, DDQs, and compliance sign-off follow a set process so onboarding does not stall in procurement. Data compliance details.
Most of the last decade building the search and recommendation systems behind products a lot of people use every day, which is the same problem this company solves: take messy real-world media at scale, verify it, rank it, and serve it accurately.
- Early Instagram engineer. First backend and ranking engineer on Instagram Search and the discovery page, co-designed the ranking behind Stories in 2016, and built the search infrastructure, applied ML stack, and experimentation systems behind a product that sees over a billion clicks a day.
- Engineering Manager at Coinbase, running the recommendations org and leading search and content ranking from zero to hundreds of millions in revenue. Coinbase engineering blog.
- Engineering Manager at TikTok across video content understanding, recommendation systems, and search ranking, driving creator monetization.
- Wharton MBA. Writing in Harvard Business Review and Business Insider, and an open-source project with millions of downloads.
- Advisor to Series C and D AI startups and hedge funds, including companies backed by OpenAI.
Anthony leads go-to-market and commercial partnerships, working directly with hedge funds, research teams, and platform partners. Over 15 years he has scaled revenue and operations at high-growth companies, including strategic finance at Tesla (SolarCity), where he led financial planning for a $6B+ bookings organization, and revenue leadership at venture-backed Betterworks and Wrike. He started his career at Equilar analyzing SEC filings data for institutional clients and media partners including Bloomberg, The Wall Street Journal, and The New York Times. Wharton MBA.
What you can do with it
Browse any executive's verified history, open a single quote to see its source video and context, or run the whole archive through the API. With the API and a model like Claude you can trace an executive's position on a topic across years, or find where they contradicted themselves.