Cal AI hit 15 million users in under 18 months.
Built by two teenagers. No VC playbook. Acquired by MyFitnessPal in March 2026 at roughly $30M ARR.
And the growth engine wasn't paid ads. It was an army of UGC creators, hired and managed like a sales pipeline.
Jake Castillo led marketing at Cal AI through that run. He just sat down with Starter Story Build and walked through the exact playbook: how he finds creators, how much he pays them, and the two-question framework that decides what every video looks like.
I watched the whole thing so you don't have to. Here's the full breakdown, with the numbers.
Why hire UGC creators instead of influencers
Jake's core line from the video: "Good outputs start with great inputs, and creators are your inputs."
Your script doesn't matter if the person delivering it is flat. The creator is the input everything else depends on, so most of the work happens before anyone films anything.
The problem: real influencers got expensive. AI made building apps easy, so everyone launched one, and everyone is bidding on the same fitness and productivity accounts. If you're a solo founder, you'll get outbid.
UGC creators are the workaround. These are everyday people, not personalities with an audience. They spin up a brand new account dedicated to your product and post for you. You're not renting their followers. You're paying for raw reach from the algorithm itself.
That's how Cal AI did volume. Not three big influencer deals a month. Dozens of small creators posting constantly. This is the same system I broke down in the Cal AI founder story, seen this time from the inside, from the guy who ran it.
Step 1: Where to find UGC creators (and what to pay)
Two sourcing channels, per Jake:
→ Post a job on a UGC marketplace like SideShift. You put up a brief and get hundreds of applicants overnight.
→ Search "UGC creator" directly on TikTok and Instagram. Creators tag themselves because they want brand deals.
Finding people is the easy part now. The filter is what saves you money.
Jake's filter, in order:
→ Talking-head content only. It converts better than anything else for apps.
→ Natural camera presence. The vibe you want: someone FaceTiming a friend, not someone reading a script.
→ Proof of reach: at least one talking-head video with 1 million or more views. One. If they've never broken a million, you're gambling on an unproven input.
Then the deal structure, and I love how cheap this starts:
→ Test campaign: 10 videos at $10 per video. That's $100 to find out if a creator can move installs.
→ If the test works: 30 videos a month at $30 per video, so roughly $900 a month per creator.
→ Performance bonuses at 100K, 500K, and 1M views, so the upside sits on the creator's side of the table too.
One more tactic that sounds small and isn't: get every promising creator on a phone call immediately. Jake is blunt about this. Negotiations that drag over email for a week close in ten minutes on a call.
And treat the whole thing like a pipeline, not a one-off gig post. Source daily, keep a list, follow up on warm creators the same way you'd run lead nurturing for your SaaS. By mid-run Cal AI had over 150 creators on retainer. That doesn't happen from one job post.
Step 2: Warm up the account before a single branded post
This is the step almost everyone skips, and it's why most founders conclude "UGC doesn't work."
TikTok and Instagram actively suppress accounts that look like dedicated ad accounts. Fresh account, first post is a product plug? The algorithm buries it. Zero reach, and you'll blame the creator or the app when the account was dead on arrival.
Jake's warm-up protocol for every new creator account:
→ The creator spends about 30 minutes a day inside your niche. Watching, liking, commenting. Behaving like a human, because the algorithm is watching for exactly that.
→ They post a few videos in the niche with zero mention of your brand.
→ You wait until one of those videos breaks 1,000 views.
That 1,000-view threshold is the green light. It proves the account isn't shadowbanned and can actually reach the feed. Only then does branded content start.
Bonus: the warm-up doubles as a work-ethic filter. A creator who won't do 30 minutes a day of unpaid warm-up was never going to deliver 30 videos a month anyway. You find out for $0 instead of $900.
Step 3: The two-question framework that decides your content
This is the most valuable part of the video. Before any creator films anything, Jake asks two questions about the app.
Question 1: Does your app have a magic moment?
A magic moment is a feature so visually impressive that watching it triggers an immediate "I need to try that." Cal AI's was pointing a camera at food and watching calories appear.
If you have one, you optimize for views. The creator builds the most viral story they can, and the app shows up naturally inside the story. The magic moment sells itself. You accept a lower conversion rate because the reach is so much bigger that it more than pays for itself.
Question 2: Does your app solve a painful problem?
If yes, you optimize for intent instead. The creator spends the first half of the video making the problem feel brutally real, so the right viewer thinks "that's exactly me." Then the app enters as the fix. Fewer views, way higher conversion.
And if the answer to both questions is no?
Jake's answer: stop hiring creators and go fix your product or your positioning. No content format rescues an app with no magic moment and no painful problem.
Buried in this section is the best product insight in the whole interview, which Jake credits to Cal AI co-founder Blake Anderson: start with the marketing angle first, then build the product backward from it. Figure out the content that would go viral, then build the thing that belongs in that content. Most of us build for months and then go hunting for an angle. Cal AI did the reverse, and it shows in the numbers.
Step 4: Brief with guidelines, not scripts
Founders love writing scripts. Jake says don't, unless you're an expert creator yourself (you're not, no shade, neither am I).
Instead, hand every creator a one-page brief:
→ The format: talking-head.
→ The framework: magic moment or problem-to-solution, per the step above.
→ General guardrails on tone and the call to action.
Then get out of the way. The creator knows what plays on their feed better than you do. You're paying for their instincts, so let the instincts run inside your structure.
Once a few organic videos prove themselves, the winners become your paid ads. You're no longer guessing what creative converts. The algorithm already told you, for a fraction of what an agency would charge to find out.
What this looks like at your scale
You're probably not trying to get acquired by MyFitnessPal next quarter. Jake frames the realistic target for a solo founder: zero to $10K a month in revenue on UGC alone, without raising money and without a real ad budget.
Run the math on the starter version:
→ 3 creators on $100 test campaigns: $300.
→ Keep the best one at 30 videos a month: $900.
→ Total first-month spend: about $1,200 for 40 shots at the algorithm.
Compare that to one mid-tier influencer post, which runs $1,500 to $5,000 for a single shot that disappears in 48 hours.
The UGC route is slower and messier. It's also repeatable, and it compounds: every video teaches you which hooks work, and the winners graduate into paid ads. If you want the deeper version of turning one app into a content machine, I broke that down in how to market a viral app with UGC.
FAQ
How much does it cost to hire UGC creators?
Cal AI's structure: $10 per video for a 10-video test ($100 per creator), then $30 per video at 30 videos a month ($900 monthly) for creators who prove out, plus view bonuses at 100K, 500K, and 1M. That's far below agency UGC rates, which often start at $150 to $250 per video, because you're hiring unproven creators and letting the test filter them.
Where do I find UGC creators for my app?
Two places: UGC marketplaces like SideShift, where one job post pulls hundreds of applicants overnight, and searching "UGC creator" on TikTok or Instagram directly. Filter for talking-head creators with natural delivery and at least one video past 1 million views before you pay anyone.
Why are my UGC videos getting zero views?
Most likely the account was never warmed up. TikTok and Instagram suppress fresh accounts that immediately post branded content. Have the creator engage in your niche for 30 minutes a day, post a few non-branded videos, and wait for one to pass 1,000 views before any product content goes up.
Do UGC creators work for SaaS, or just consumer apps?
The framework transfers: if your SaaS has a visual magic moment, optimize for views; if it solves a painful problem, optimize for intent with problem-first videos. Consumer apps convert faster from TikTok, but B2B founders are using the same talking-head format on LinkedIn and X with smaller numbers and higher deal values.
How many videos before I know if UGC is working?
Jake's test is 10 videos per creator. Across 3 creators that's 30 data points, enough to see whether anything breaks 1,000+ views and drives installs. Judge the system after a month of consistent posting, not after two videos.
Steal the playbook, then hear the next one
Every week on the Profitable Founder Podcast I pull apart stories exactly like this one: real founders, real numbers, playbooks you can run the same week.
If Cal AI's UGC engine taught you something, the next episode will too.
