Growth Strategy · Reforge Foundation
Product-Market Fit Collapse
PMF is not permanent. AI is causing sudden, complete PMF collapse in categories that previously had stable demand. Brian Balfour's framework for diagnosing collapse risk — and what to do about it.
PMF is a state, not a destination
The conventional narrative around product-market fit treats it as something you achieve and then keep. Find PMF, scale on top of it, defend the moat. Brian Balfour's argument in Product-Market Fit Collapse is more uncomfortable: PMF is a state, not a destination. It can be lost. It can be lost suddenly. And AI is causing it to be lost in categories that previously felt stable.
The collapse pattern is specific. It's not slow decay (that's the normal lifecycle of any product). It's sudden, large-magnitude loss of fit because the market itself has moved in a way that makes the product's value proposition obsolete. Homework-tutoring services lost most of their PMF in months when ChatGPT made the core use case zero-cost. Copywriting agencies that built on $0.10/word freelancers face the same dynamic. Sub-$10/month productivity SaaS products face the question of whether AI-native alternatives have made their core feature set table-stakes.
Why AI accelerates collapse
Previous technology shifts caused gradual displacement — incumbents had time to adapt because adoption was slow, expensive, or required infrastructure. AI is different in three ways that accelerate collapse:
- Adoption is fast. No installation, no procurement cycle, no infrastructure investment. A user can switch from a paid SaaS tool to a ChatGPT prompt in seconds.
- Cost is near-zero at consumer scale. ChatGPT free tier, Claude free tier, Gemini free tier. Most consumers and many small businesses can replace paid tools with no marginal cost.
- Capability ramps fast. Models improved 10x in capability between 2022 and 2025. Use cases that were nonviable in 2022 became dominant in 2024.
The combination produces collapse curves that are steeper than anything previously seen in software.
The diagnostic — are you at risk?
The Reforge framework offers a four-factor diagnostic:
- Use Case risk. Can the core user task be done with a general-purpose AI tool? Homework: yes. Brain surgery: no.
- Growth Model risk. Does your acquisition depend on channels that AI is reshaping? SEO content sites face AI-Overview risk. Paid social faces creative-cost compression.
- Defensibility risk. What makes you hard to replace? Brand, integrations, data moats, network effects — or just a UI on top of capability that's now commoditized?
- Business Model risk. Does your monetization assume features that AI is making free? Per-seat subscription on a capability that's now consumer-grade?
High risk on any one factor is manageable. High risk on three or four factors is collapse exposure.
What to do about it
The Reforge prescription is uncomfortable: lean in. Don't defend, build the AI-native version. Don't wait for collapse, accept that current PMF is decaying and build the next state actively.
Reforge itself did this — they shipped five AI-native products with a team of 20, including Reforge Build (their AI prototyping tool). The bet was not "AI will displace some of our content business" but "AI will displace all of it unless we build the AI-native education company first."
RGM experts say
The single most dangerous failure mode for companies facing AI disruption is denial. We have worked with operators who, six months into measurable revenue decline correlated with ChatGPT release dates, were still framing the issue as "marketing optimization" rather than "PMF collapse."
The diagnostic question that cuts through denial: if a competent solo operator with ChatGPT access could replicate 70% of your product's value in a weekend, you are exposed. Acknowledge it and build the next state.