In China, AI Can't Sell Tools — It Must Deliver Results: Even Sam Altman Would Leave Battered｜A Conversation with Zhang Shaofeng, Founder of Bairong AI
The AI era is shaping up to be a tragedy for middle management. If agents can handle every relay of information up and down the chain, why keep a permanent layer of people whose only skills are running meetings and filling spreadsheets?
Take Square, the company Twitter's former founder built. Earlier this year they axed 40% of their staff. The founder laid it out bluntly: the modern corporate hierarchy traces straight back to the Roman legions! One person can only manage five to eight others, so you need decurions and centurions. What do these middle managers actually do? They're human loudspeakers. If AI relays messages more accurately, faster, and without social security contributions, what's the point of permanent middle management?
My Silicon Valley friends don't greet each other with "have you eaten?" anymore — it's "got laid off again?" To protect themselves, everyone's collecting partner titles at AI startups. The funniest part is the interview gauntlet: one hour on the clock, first thirty minutes you're free to blast away with Claude and Cursor, watching you finish a week's work in thirty minutes. Then the interviewer bursts in, yanks your ethernet cable, kills your AI access, and sneers: "Now hand-write the rest."
Why so sadistic? Because capitalists aren't stupid! They're not just testing your "vibe coding" skills — they need to see if you can clean up the mess when AI shits the bed. Zuckerberg said it himself the other day: eventually they'll fire most developers, because their entire workflow is being used to train the AI that replaces them.
AI Babysitters — You Think That Salary's Free?
From "35 Minutes of Getting Ripped Apart" to "Our Only AI-Native Company"
Think the entrepreneurs are anxious? The investors are worse. Don't kid yourself. These FOMO-screaming VCs are privately terrified — terrified they backed a portfolio of scrap metal, terrified of missing the high-speed rail to the future.
Zhang Shaofeng, founder of Bairong AI, told me a surreal story. The year before ChatGPT exploded (2021), his investor summoned him and tore into him nonstop for 35 minutes. Why? The investor was forcing him to pivot to the American SaaS model. Back then SaaS was catnip for capital markets, with trillion-dollar valuations floating around — companies like Weimob and Youzan were riding sky-high.
The investor sprayed saliva everywhere. Shaofeng sat in silence. When the investor finally ran out of steam, he fired back coldly: "SaaS doesn't work in China! Chinese enterprises will never pay for process!"
Sure enough, two years later, SaaS valuations got their ankles sliced — down 60-70%. Anyone still pumping SaaS has brain damage. The kicker? When the ChatGPT hurricane hit, the same investors came running: "Shaofeng, what's our AI pivot?" Upon learning Bairong had already deployed agents and RaaS (Result as a Service), they instantly flipped: "Bairong is the only AI-native company in our entire portfolio!"
The picture writes itself.
Even Sam Altman Would Cough Up Three Liters of Blood
Why does SaaS fail in China? This is the part that makes me want to scream.
Try selling SaaS to a Chinese boss and it's like persuading a landlord used to hiring farmhands to buy a high-tech sickle. The landlord's response? "Nice sickle, but I don't know how to use it! Why don't you stay and harvest this field for me, and we'll split the profits 50-50?"
Chinese enterprises only recognize two things: either you bring me resources and traffic, or you do the damn work and deliver results to my desk. You want me to pay for your software tool, then hire people to learn it? Dream on!
Plenty of AI startups march into big tech with slick agent demos. What happens? The bosses wave the "digital transformation" flag, milk your free consulting to map out their business logic, then dismiss you with: "Hmm, this product isn't mature enough." Next thing you know, your proposal's been handed to their internal tech team.
I was in SaaS too — we ended up hanging banners: "Free consulting for business opportunities! The revolution continues, comrades must persist!" That's naked Chinese B2B reality. You think you're driving AI transformation; they treat you as free external brainpower!
If you parachuted Sam Altman or Anthropic CEO Dario Amodei into China to do B2B sales, I guarantee within three days these Silicon Valley titans would be coughing blood. In America, you build a tool, sell by the seat, people pay up. In China, the boss only asks: "Can your AI move ten more crates of product? No? Get lost!" Chinese SaaS never truly lived — it's all zombies. Chinese people buy houses, don't rent; they'll always be at war with the landlord!
Even the former boss of a telecom call-center software company actively sought acquisition by Bairong. Why? They used to charge by "human seat count" — now AI vaporized the humans! Seat count hits zero, and who cares how high your unit price is? That's the final destination of traditional SaaS! AI isn't building a better shovel; AI is eliminating the person wielding the shovel.
Results Are the Only King
So stop babbling about SaaS. OpenClaw has already ripped off that fig leaf: SaaS is dead, the next wave is RaaS!
AI isn't a faster screwdriver — AI is the employee turning the screw. If you can tackle long-horizon tasks that outsourcing can't touch, if you can deploy full-time digital workers to slash costs and deliver results directly, you get paid. The client boss doesn't care if it's done by human or AI — as long as you get things done, as long as costs are lower than before, they'll sign.
But this creates a lethal problem: to deliver results, you need to understand the business! You need know-how! And what's the reality of Chinese enterprises? They don't even have proper knowledge bases!
I visited one company to build AI customer service for them. Turned out all core knowledge lived in the heads of a few "old masters." Ask them to write SOPs? Impossible. First, I'm busy meeting clients; second, teach the apprentice and the master starves — why should I teach you? Your digital foundation is rotten to begin with, how do you do RAG (retrieval-augmented generation)? It's like building a skyscraper in mud. Garbage in, garbage out!
The Great Shakeup Has Begun — No One Stays Untouched
Today, every enterprise faces internal and external pressure. Economic headwinds, harder hiring, traditional outsourcing models hitting a dead end.
Theoretically, Chinese enterprises need Palantir-style deep-customization data services far more than America does. But thanks to China's toxic payment habits, the Palantir model died here long ago. Only RaaS and FDEs (full-time digital employees) offer a way out. The advantages of silicon-based staff are brutal: no social security, no emotional outbursts, no PIP cycles, free trial for months, pull the plug and terminate anytime — no severance either.
Shaofeng told me when a deputy president at a major bank proactively messaged "I might consider hiring your silicon employees," he knew the wind had shifted completely. This is no longer about tool upgrades — in their telemarketing business, AI took over the cost center entirely. Using AI to chat, negotiate, and close deals, putting quantified conversion results directly in front of the boss. In this process, we're no longer in a "software vendor" relationship with clients — we're profit-sharing partners. Bairong helps clients sell, then we split the take, commission-style.
On this global AI battleground where China and America are separated by a cliff-like gap, any company still numbing itself with "SaaS dreams," fantasizing about licensing software and collecting rent lying down, is committing slow suicide. China's market complexity and extreme utilitarianism make it the best training ground for AI — and the most brutal meat grinder.
In China, don't sell screwdrivers — go turn the screws yourself!
【Subscribe to JustSayAI Daily Brief】 🌈 New here! Subscribe to the「JustSayAI Daily Brief」· Two editions daily · One-click audio 🌟 Returning members! Join membership for JustSayAI Daily Brief + deep-dive research + written columns
【Host on Xiaohongshu】 📕Call Me Xiao Su
【Follow the Channel】 ▶️YouTube｜ 📺Bilibili｜ 📕Xiaohongshu (RedNote)｜ 🎵Douyin｜ 📻Podcast 📱 Business inquiries/listener community: justsayai666