How a seed-stage AI agent startup got 1,840 activated workspaces in 8 weeks

A YC-backed, seed-stage AI agent startup got 1,840 activated workspaces in 8 weeks from 9 YouTube creators, 6 July to 30 August 2026. The decision that drove it was briefing each creator to run the agent on one real business task, start to finish, on camera: full-run videos activated 44% of their sign-ups against 33% for feature tours.

Written by Emma Park

Prefer a conversation? Book a call

The company and program

ClientA YC-backed, seed-stage AI agent startup.
StageSeed, YC-backed
IndustryAI agents
Timeline8 weeks, 6 July – 30 August 2026. First video live on day 16
ChannelsYouTube: 3 dedicated videos and 10 sponsored integrations from 9 creators
Team1 dedicated Kiko growth lead, plus agents for creator scoring, script accuracy checks and daily link reporting

The approach

  • Hypothesis: small-business owners sign up and activate when they watch the agent finish a real task they already do by hand, not when they watch a feature tour.
  • Comparison: full-run videos (one real task, start to finish, on screen) against feature-tour integrations, across the same creator groups. Cost per activated workspace against the founders' LinkedIn ads from the 8 weeks before.
  • Decision rule: after 7 days of organic data on each video, rebook any creator under $40 per activated workspace. Drop any creator over $80.

What we changed

  1. Weeks 1–2: score and sign. We scored 64 channels on Kiko's six-point rubric, shortlisted 14 and signed 9. The roster: three small-business automation YouTubers at 150K–450K subscribers, two agent-builder tutorial channels at 90K–250K, two AI-for-operators analysis channels at 200K–400K and two AI agency-owner channels at 80K–200K.
  2. Week 2: brief. One page per creator: pick one task the viewer does every week (chasing unpaid invoices, triaging a shared inbox, updating a CRM), connect the tool live, and let the run finish on screen. Kiko checked every script for feature accuracy and FTC disclosure.
  3. Weeks 3–5: first wave. Nine videos went live with tracked links and a free-month code, and the sign-up form asked "Where did you hear about us?".
  4. Week 6: read and rebook. After 7 days of data per video, five creators came in under $40 per activated workspace; we rebooked the four with open slots and dropped promo codes from the new briefs.
  5. Weeks 6–8: second wave. Four more videos went live, and the client added a "start with the task you saw" onboarding template keyed to each creator's link.
Kiko program report showing the before and after results across the full program. Recreated from the case-study results; account details omitted.
Kiko program report showing the before and after results across the full program. Recreated from the case-study results; account details omitted. View full report

The results

MetricBefore (prior 8 weeks)After (weeks 1–8)Change
Sign-ups, all sources1,6806,320 (4,600 from creator links)3.8x
Activation rate (3 successful runs in 7 days)24%40% for creator sign-ups+16 pts
Activated workspaces4031,840 from creators, plus 430 from other sources5.6x all sources
Cost per activated workspace$118 (LinkedIn ads)$30.00 (creator fees)−75%
Activated workspaces on a paid plan by day 3014%17% (313 workspaces)+3 pts
Full-program results in PostHog’s dashboard layout, including the acquisition and conversion figures from the results table. Recreated from the case-study results; account details omitted.
Full-program results in PostHog’s dashboard layout, including the acquisition and conversion figures from the results table. Recreated from the case-study results; account details omitted. View full report

What didn't work

  • Analysis channels drove views, not activations. The two operator-analysis creators delivered 616K views but cost $40 and $56 per activated workspace. Their viewers wanted to understand agents, not install one this week.
  • Promo codes added nothing. Only 9% of creator sign-ups used the free-month code; the rest came through the tracked link. We dropped codes from the second wave.
  • A live run failed on camera. One agent run timed out on an accounting sync mid-recording. The creator reshot it, and the video went live a week late.

What we learned

  • Brief one real task per video and let the run finish on screen. Full-run videos activated 44% of their sign-ups; feature tours activated 33%.
  • Rebook winners after 7 days of data instead of booking the whole quarter up front. All four rebooked creators beat their first video.
  • Match onboarding to the video. Starting new users on the task they just watched lifted tool connections in weeks 6–8.

Frequently asked questions

How much does a YouTube creator campaign for an AI agent startup cost?

Each activated workspace cost $30.00, against $118 on the founders' LinkedIn ads, with no paid media. Most seed programs like this run $15,000–$35,000 a month in creator fees and media.

How long until creator videos drive sign-ups?

The first video went live on day 16; nine ran in weeks 3–5 and four rebooked videos in weeks 6–8. Each got 7 days of organic data before the rebook call. The 13 videos drew 2.86M views and 4,600 sign-ups.

Would this work for a Series A AI agent platform selling to enterprise teams?

The full-run brief carries over; the audience doesn't. These creators reach small-business owners who install a tool the same week, and even the operator-analysis channels here cost $40–$56 per activation. For enterprise buyers, judge creators on demos, not activations.

Plan a creator campaign for your product

Prefer a conversation? Book a call