How Jordan Gal pivoted to an AI phone answering service with hockey stick growth

Executive overview

Most SaaS pivots burn cash and lose momentum. Jordan Gal shut down a struggling e-commerce checkout, cut his team from 20 to 6, and shipped an AI voice answering service for small businesses in under 90 days.

The result: doubling MRR every month, approaching $1M ARR, and churn far below initial projections.

Replacing voicemail — not a human — lowers the bar enough to win fast.

From Cart Hook to Rosie: the pivot story

• Cart Hook grew from $1M to $6M ARR organically, then Shopify killed third-party checkouts
• Follow-up company Rally (headless e-commerce checkout) stalled at a few hundred K ARR
• A conversation with Justin McGill — experiencing product market fit — reminded Jordan what growth is supposed to feel like
• Team reduced from 20 to 6; those who stayed chose to stay
• Idea search landed on AI voice agents for small businesses; first paying customers within ~3 months

What product market fit actually felt like

• Qualitative first: frantic support chat, customers begging for it to work, asking to jump on calls
• Quantitative later: hockey-stick graphs, screenshots posted in disbelief
• Initial churn assumption: 15–20% monthly; actual churn: far lower
• Key insight: Rosie doesn't replace a human receptionist — it replaces voicemail, so the bar to deliver ROI is low
• The market poll feeling — customers pulling at you, demanding, not needing to be convinced

The accordion feature strategy

• Launched with ~8 features, intentionally broad, to discover what users actually needed
• Within weeks, 2–3 features emerged as essential; the rest were nice-to-haves
• Removed non-essential features from the UI entirely (kept in code, just hidden)
• Core feature set became the $49/month base plan: answering questions and taking messages
• Nice-to-haves (call transfers, appointment setting, CRM integration) reintroduced as higher tiers once better understood
• Individual features have their own URLs in the admin — easy to toggle per user

Building fast as an AI wrapper

• AI products require far less custom build than traditional SaaS — the infrastructure (LLM, voice, transcription) is leased, not built
• Team went from strict QA and deployment processes (required for a checkout) to intentionally burning 90% of that process
• Speed came from: smaller team, simpler product, deliberately discarding old habits
• "AI wrapper" criticism misses the point — DeepSeek commoditised the LLM layer, which shifted value back to the application layer
• Land grab logic: get customers in while the market is forming; they won't switch if the product works

Onboarding as the growth inflection point

• Deliberately delayed building onboarding until the base feature set was locked
• Primary onboarding path: enter Google Business Profile → AI pulls structured data → plays a demo call with the business's own name in under 3 minutes
• Credit card placed after the demo call, not before — users put it in voluntarily because they don't want to lose access
• Inspired by Justin McGill: drag value out of the admin into the marketing/signup experience
• Non-technical users (painters, electricians) fully onboarded on a phone while sitting in a truck

Switching from time-based to usage-based trials

• Original trial: 7-day with credit card required — looked great initially, then discovered low usage underneath
• Switched to a 25-minute talk-time trial (usage-based, not time-based)
• Result: everyone who converts is an activated user; churn dropped significantly
• Lesson from Cart Hook: a 15% monthly churn "washing machine" caps your MRR ceiling regardless of growth rate
• Healthier cohort > faster revenue; tested as a 30-day experiment before committing

Pricing, ARPU expansion, and what's next

• Three tiers: $49, $99, $199/month
• COGS are tied linearly to minute usage — unlike typical SaaS, not negligible
• Actively moving away from "minutes" framing toward value framing
• Premium UI launched: yellow stars on locked features in the admin reveal upgrades in a self-serve flow
• Nearly 50% of revenue now comes from higher tiers since premium UI launched
• Horizontal positioning (all small businesses, no vertical niche) deliberate — mirrors early email marketing tools like Constant Contact, AWeber
• Mindset: "extreme optimism and paranoia" — push for $10M ARR in 18 months while expecting disruption at any moment