BetterPic.
From $1K to $270K MRR in 17 months. A four-person team, an AI-assisted stack, and the receipts to show for it.
Monthly Recurring Revenue, Jan '24 → Jul '25
Data approximated to nearest $2k · self-reported · independently verifiable on request
The situation
I joined BetterPic in late 2023 as the first growth hire. Revenue was around $1,000 MRR. The product (AI-generated professional headshots) worked. Acquisition was organic-only and the funnel had no quantitative spine.
The mandate was simple. Take a Seed round and turn it into a growth engine that justified a Series A. The honest version: we had a window to prove that paid acquisition could work in a category most investors had written off as a feature waiting to get absorbed.
The playbook
We didn't invent a new playbook. We compressed an existing one and forced every step of it to be AI-native.
- Re-instrument the funnel. Every step measurable, every step a hypothesis.
- Land one creative format that worked. Then industrialize the testing loop with AI-assisted variation.
- Layer in conversion lifts (CRO, onboarding) once paid was predictable.
- Add compounding channels (SEO, lifecycle, community) once unit economics held.
Channel breakdown
At peak month, the mix looked roughly like this. Numbers approximate. The point is the shape of the engine, not exact percentages.
LLM-assisted creative iteration cycle. Roughly 40 net-new concepts per week. Sustained 4× ROI at peak.
Brand and high-intent keywords. Tight CPA discipline, automated bid scripts.
Comparison pages and listicles built to rank in both Google and LLM citations.
Real, in-the-trenches participation in relevant subreddits. Conversion rates that paid social can't touch.
Demo content and tutorial-led growth. Long-tail discovery for both B2C and B2B.
Lifted account creation and purchases 1.6× from existing traffic with a re-architected onboarding flow.
Where AI did the work
"AI-native" was load-bearing here, not decorative. Three places it mattered most:
- Creative iteration. A creative-ops loop that generated roughly 40 net-new ad concepts per week with one human in the loop. Without it, the testing cadence wouldn't have been possible.
- Landing page personalization. Per-ad-set landing pages auto-generated and lightly QA'd by a human. CVR lifted measurably for the segments where it mattered.
- Inbound research. Agent-driven SERP and LLM citation research so the SEO team was always ahead of competitor comparison pages.
What didn't work
Some things didn't survive contact with the market. Listing them because pretending they did would insult anyone who's actually scaled a SaaS company.
- Influencer-led acquisition. Looked obvious for a visual product. CAC was inconsistent and attribution was a mess at our scale. We paused after a quarter.
- B2B-only pivot tests. Spent weeks trying to sell exclusively into HR teams. The conversion economics weren't there at our ACVs. We returned to a consumer-led motion with B2B layered on top.
- Heavy outbound and SDRs. Net negative on margin at the price points we operated at.
Takeaways
- The funnel comes first, the brand comes second. Brand comes after paid becomes predictable. In that order.
- AI compresses cycle time. The creative concepts still came from humans. AI made the testing loop roughly 10× faster.
- One channel, then layer. Trying to be on five channels at $1k MRR is the most common failure mode I see.
- Hire for taste, not headcount. Three operators with an AI stack beat a team of fifteen on this kind of motion.
If you're scaling a SaaS startup and want a second set of eyes on the growth motion,