AI Monetization: Treat Users as Customers, Not Gods
AI Summary · Perspective of a Serial Entrepreneur
By reviewing their own AI spending habits, the author divides AI users into four payment tiers: Free users benchmarked against WeChat/Meituan (zero willingness to pay); Per-use benchmarked against video subscriptions (<100 RMB for a smooth experience); Monthly subscription benchmarked against hourly contractors (100–200 RMB to buy time); and Pro users benchmarked against production equipment (>200 RMB calculated by ROI). The core takeaway: willingness to pay is determined not by income, but by the intersection of “service value” and “ease of use.” Entrepreneurs should design tiered products accordingly rather than blindly chasing ARPU.
I’ve spent years grinding in the startup world, watching too many products fail because they focused on “how to get users to pay” instead of “figuring out what users are actually paying for.” This piece hits on an overlooked truth: AI monetization isn’t a pricing problem—it’s a matching problem.
1. Free Users Aren’t Stingy—They Just Haven’t Found the Right Hook
Doubao, Kimi free tier, DeepSeek free tier… these people use AI the way they use WeChat. Have you ever paid to use WeChat? No. Because chatting is a free infrastructure layer on the internet.
The key insight: free users’ payment threshold is “something concrete and perceptible.” You pay for video subscriptions to watch shows, for game skins to upgrade your look, for a rocket emoji to hear a streamer say thank you. But what did you buy with an AI chat? Nothing tangible.
→ Takeaway for founders: Don’t expect free users to pay for “intelligence” itself—pay them for “mental bandwidth saved.” Package AI as “auto-categorizing your to-do list” or “pushing your top priorities every morning,” not as “a smarter assistant.”
2. 100 RMB Is the Mental Tipping Point
People willing to pay for AI assistance typically budget under 100 RMB. Why? Because that’s what it benchmarks against: iQiyi, Tencent Video—skip the ads, watch in HD, enjoy a smoother experience. Push past 100, and users start hesitating: “Why should I pay three platforms for one membership?”
→ Takeaway for founders: If your product only delivers a “smoother” experience, price it under 100. To break past that ceiling, you must elevate from “more convenient” to “does work for me.”
3. At 145 RMB, You’re Hiring an “Hourly Contractor”—Not Buying a Tool
Claude Pro and ChatGPT Plus can charge 145–200 RMB because users treat AI like an hourly contractor: on-call, no meals required, never quits. At that point, the payment logic shifts from “buying something” to “paying a salary.”
→ Takeaway for founders: To command 200+ RMB, you must prove the AI is genuinely helping them work every day—and that the time saved is worth the price. Otherwise, users will think: “I’m paying for an assistant, and it’s just chatting with me?”
4. Pro Users Don’t Look at Price—They Look at ROI
For fifth- and sixth-tier users (Vibecoding enthusiasts, developers), AI is production equipment. They blast through APIs because the math works: 20 hours saved × 200 RMB/hour = 4,000 RMB return, minus 300 RMB spent = 13x ROI.
→ Takeaway for founders: These are your lighthouse customers. Serve them well—their case studies become your best referrals.
5. The Barrier Sets the Payment Ceiling
With each tier up, the barrier rises: free → API-literate → fluent in complex tools → knows how to code. Most people get stuck at the 100-RMB level.
→ Takeaway for founders: Product tiers must map to capability tiers. To break past 100 RMB, you must either lower the operational barrier to “getting things done” or raise the reliability of the output.
Final Thoughts
What resonated most with me is that the author doesn’t assume “users don’t understand value.” Instead, they蹲 down to ask: “What’s actually inside the user’s mind?” As founders, we often make the mistake of thinking users ought to see our product’s value. The truth is, users only pay for what they’ve internally benchmarked.
Next time someone asks “How should AI be priced?” don’t start with cost—ask: Does the user see AI as WeChat, an hourly contractor, or production equipment?
Original article · Everyone's a Product Manager: Read original →