Notes from the team rebuilding around AI.
Field notes, product updates, and worked examples from the team building ThoughtLeaders.AI. New posts on launch days and around the major releases.
Your break-even CPM: the most you should pay for a YouTube sponsorship
Stop asking whether a creator's quote is fair and start asking whether you can afford it. Two numbers from your own past campaigns give you a break-even cost per 1,000 views. From there, every quote becomes a walk-away price you can work out in seconds.
Read the full postEarlier posts
YouTube sponsorship cost by niche: what brands paid per 1,000 views in 19 niches
Beauty, business and finance channels sold at two to six times the price per projected view of news and commentary channels, and channel size explains much of that gap. Here is what brands paid per 1,000 projected views in 19 YouTube niches, across 9,098 sponsorships booked through ThoughtLeaders, adjusted for channel size, plus what YouTube's August view count change does to those numbers.
We asked seven AI models, including GPT-6 Astra, Claude Opus 5.5 and Gemini 4, what a YouTube sponsorship costs. For big channels, even their low end was 2 to 10 times what brands paid
Give an AI model a subscriber count and it overprices big YouTube channels. For a 2 million subscriber channel, the bottom of the typical answer was $20,000 from the models behind ChatGPT, Claude, Gemini and Perplexity, and $6,500 from the strongest models (GPT-6 Astra, Claude Opus 5.5, Gemini 4 Argon), while brands paid a median $3,500 through ThoughtLeaders (data as of October 2026). Give it views instead and every model landed close to real prices. We ran 15 buyer pricing prompts through seven models' APIs, twice each, and set all 210 answers against what brands actually paid.
Dedicated YouTube video vs integration: channels ask 2x, brands paid 1.33x
Channels list a dedicated YouTube video at twice their integration price, but matched dedicated deals bought through ThoughtLeaders since 2024 sold for a median 1.33 times the same channel's integration, and almost a third at no premium. The likeliest route to that discount is a channel where a whole video about the product is content the creator would make anyway. Here is the price data, why creators resist dedicated videos, how they renew and perform, and how to find the right fit.
Competitor sponsor gap audit: the YouTube niches your rival keeps renewing that you barely tested
Over the past year LMNT renewed 35 health and fitness YouTube channels with a separate booking, while Liquid I.V., which sells a very similar electrolyte mix, renewed 3. Here is the niche-by-niche gap table for the two brands, what it shows, and how to run the same audit for your own brand with your AI agent.
Before you book a YouTube channel, check which sponsors came back
A channel's sponsor list tells you who tried it. Its renewal rate tells you who came back, and nobody renews a channel on vibes. Here is what a typical renewal rate looks like across 37,266 sponsored YouTube channels, and how to read one before you book.
YouTube changed how it counts views. Here is what that does to your sponsorship math
Since August 24, 2026, a YouTube view counts from the first frame, so long-form view counts look bigger and every sponsorship CPM worked out on them looks cheaper. Here is what changed, how much it moved by day 30 in our measurement, and how to compare channels, quotes and guarantees on the same basis, with a real run. Updated October 27, 2026 with day-30 data.
Found a YouTube channel that works for your brand? Find more channels like it with your AI agent
One channel that worked is the best brief you will ever write for your next YouTube buy. Here is how to turn it into a shortlist of look-alike channels with your AI agent, filter out the ones a sponsor can't use, and vet the rest on YouTube, with a real run.
YouTube sponsorship market: October 2026 update
One in 16 new long-form YouTube videos we scanned carries a detected sponsorship, up from one in 21 a year ago. Our October 2026 read on the YouTube sponsorship market: volume, the top brands, who is growing fastest, the busiest categories and what sponsored channels are projected to get per video.
YouTube CPM vs RPM vs sponsorship CPM: three numbers that price different things
YouTube's CPM and RPM measure YouTube's own monetization; sponsorship CPM measures what a brand pays for a creator's own mention, and the two markets don't price the same thing. Here is how each is calculated, what brands actually paid per thousand views by channel size, and how to turn the numbers into a fair quote.
Which YouTube creators does your competitor sponsor? Find their repeat channels with your AI agent
The fastest way to find YouTube channels that work for your brand is to see which channels a similar brand keeps coming back to. Here is how to map a brand's repeat sponsorships with your AI agent, expand them into look-alikes, and vet them on YouTube, with a real run.
Which AI tools sponsor YouTube creators? Q3 2026 data
AI tools racked up 8,113 detected YouTube sponsorships in Q3 2026, up 18% on Q2, and half of them ran on tech channels. Here is who sponsored the most, which channels they keep coming back to, and how to run the same analysis for your own category with your AI agent.
How to check a YouTube sponsorship quote with your AI assistant
A creator sent you a price (in our example, a hypothetical $5,000 quote on an example channel). Before you sign, ask your AI agent to check it against the channel's projected views and a fair-value estimate, see what is driving the number, and turn the gap into a counter-offer. Here is the exact workflow, step by step.
How to check a YouTube channel for fake views before you book a sponsorship (with your AI agent)
Bought views look exactly like real ones on the view counter, and you pay for both the same way. Here is how to vet a YouTube channel for fake views and bot comments before you sign: the prompt to give your AI agent, what each signal means, how to read the verdict, and the manual checks you can run yourself on YouTube.
How much does a YouTube sponsorship cost? What brands actually paid, by channel size
Real prices from more than 9,000 YouTube integrations booked through ThoughtLeaders, broken down by channel size, with the cost per thousand projected views. Also: why YouTube's August 2026 view-count change makes price comparisons tricky.
What's new in tl v0.9.10: Claude Code plugin stays current, for real this time
If your Claude Code plugin has quietly been sitting behind the CLI for a few releases, v0.9.10 catches it up. One command — `tl setup claude` — now actually moves the plugin forward, and Claude Code picks up the freshest skills after a quick restart.
What's new in tl v0.9.9: Windows users blocked by Smart App Control get a one-command fix
If you're on a Windows machine and `tl` refused to run with a Code Integrity error, v0.9.9 has the fix: a single `tl setup windows` command sets up a signed launch path that Smart App Control is happy with — nothing to sign, nothing to disable, and your AI agent picks up right where it left off.
What's new in tl v0.9.8: when the server says no, it now says why — and what to do about it
Every refusal now comes back in plain English: which limit you hit, how much of it you've used, when it frees up, and — for a subscription org — the real menu of options instead of a one-size top-up pitch. `tl balance` shows your session-quota percentage. Data withheld for your plan tier is announced explicitly instead of showing up as silent gaps.
What's new in tl v0.9.7: a memory your agent keeps between sessions, linked reports that stay in sync, and a clearer whoami
Your AI agent can now remember the context you keep re-explaining every session — your accounts, your priorities, your naming conventions. Reports can pull entities from other reports as live linked sources instead of one-shot snapshots. And `tl whoami` finally tells you whether you're on the Industry Weekly newsletter.
What's new in tl v0.9.4: your agent finds the extra skills your org publishes, and raw-SQL detours end sooner
The extra skills your organization publishes are now discoverable to your AI agent the first time it looks for them. And two of the most common raw-SQL wrong turns now come with a helpful hint on the spot, so a curious deep-dive doesn't dead-end on a naming mismatch.
What's new in tl v0.9.3: build sourcing workflows from your AI agent, and install extra skills your org distributes
Your AI agent can now design a full sourcing funnel and drop it straight into the ThoughtLeaders app as a real workflow you can open, share, and iterate on. Plus a new skill-delivery system that installs any extras your organization publishes, and a cleaner view of the credits you have to work with.
What's new in tl v0.9.2: a protected email archive, sharper keyword research, and error messages that actually help
A tighter, safer release. Your channel email archive is now protected from accidental overwrites, keyword research reaches deeper into the data to find more matching channels, and when your AI agent hits a query error it gets a real hint instead of a dead end.
What's new in tl v0.9.0: keyword research hands you a validated filter, a channel roster, and trending videos
Ask your AI agent to research a topic — a niche, a trend, a category — and you now get back three things at once: a validated keyword filter you can save and reuse, a ranked shortlist of channels tiered by how much they actually cover it, and the trending videos on that topic sorted however you want. Plus a built-in help mode that costs nothing to ask.
What's new in tl v0.8.1: fewer wasted credits and gentler error messages for raw data queries
Your AI agent now catches itself before running the same expensive query three times in a row, and when a query does trip on a renamed field it gets a nudge in the right direction instead of a dead end. Small polish, real savings.
What's new in tl v0.8.0: clearer field names across the board and Media Selling Network editing from the CLI
A friendlier vocabulary lands across the tl CLI: advertisers, subscribers, projected views, scheduled dates — names that mean what they say. Plus, your AI agent can now add or remove a channel from the Media Selling Network in a single command.
What's new in tl v0.7.15: ask for a topic, get a ranked, validated channel shortlist
Topic-driven channel discovery just got a serious upgrade. Tell your AI agent the topic you care about — investing, cold plunges, AI cooking, whatever — and instead of a raw keyword count, it now returns a ranked, sense-checked shortlist of channels that actually cover it, with sponsorability flags ready to go.
What's new in tl v0.7.14: smarter keyword search and a sharper channel-update skill
A new switch on raw keyword search keeps the matching snippet right next to each result — so when your agent answers "which videos talk about VPNs?", you can see exactly where. Plus a clearer playbook for updating channel contacts.
What's new in tl v0.7.13: brand-to-channel matching, end to end
Two new commands close the loop on one of the questions every brand researcher asks: which creators is this brand already in love with, and who looks like them? Plus a cleaner, simpler vocabulary for sponsorship statuses across the board.
What's new in tl v0.7.12: no more ceiling on deep data sweeps
Two smooth upgrades land today: your AI agent can now walk through huge result sets end-to-end — past the old 10,000-row ceiling — when answering broad sweeping questions, and setting tl up inside the Claude desktop app just got a lot more reliable.
What's new in tl v0.7.11: smarter creator search and automatic YouTube discovery
Looking up a creator by name now works the way you'd expect — typos and odd spacing get sorted out for you, and a name we don't track yet is found on YouTube and queued for analysis on the spot.
What's new in tl, v0.7.7–v0.7.10: deeper discovery, fairer pricing, cleaner reports
Your AI agent can now pull the complete list of creators who fit a topic — not just the top few thousand — the cost structure is simpler and only charges for what you actually ask, and building or saving a report now follows one clear path.
What's new in tl v0.7.6: smarter counts, friendlier article results
Two small but useful upgrades to the tl CLI's AI guidance — your agent now answers "how many?" questions in one shot instead of paging through every record, and article searches come back with channel names already attached.
Why we bet the company on AI in 2026
This year, we made the call: rebuild the company around AI. Not a product feature. Not an internal pilot. The entire operating system of the business. Here is why, and what it has produced.