The better AI gets, the more exhausted I am
AI can write essays, design posters, and blast out a thousand emails in a single night. Yet many of the clients Coco signs were added on WeChat at a Chinese calligraphy event.
Write some characters on the spot, add a WeChat, leave a lead. That primitive.
Coco works as a sales rep at a well-known study-abroad agency. She majored in broadcasting and hosting as an undergrad, went to the UK for a master’s in linguistics, and now helps Chinese students apply to British universities.
It sounds like a mismatch, but it fits perfectly. The expression she trained through broadcasting, plus a natural warmth, happen to be the two things sales values most. She says whether you get a good mentor in this line of work makes a huge difference. Her former mentor was in the top tier; back when the market was good a few years ago, the annual income could reach several million yuan.
That is a story from another era.
Coco uses AI now too. Gathering information, organizing materials, drafting an essay. It really does save a lot of effort.
But,
“The better AI gets, the harder I have to work. The harder I work, the more exhausted I am.”
The company also knows everyone has AI in hand, so the KPIs handed down keep climbing with it. The time you save does not turn into rest. It only turns into the next, higher signing target.
The Coco who traveled all over the world as a student now just wants to lie down every weekend after starting work.
“It is not that my body is tired. It is my heart that is tired.”
Number of signings, value of signings. Every single one is a KPI pressing down on your head.
For a study-abroad agency, only one thing matters most: getting clients. It is also the only part the company is truly willing to spend money on for AI.
And yet this is exactly the part where AI cannot help.
“You can blast a thousand emails and still not close a single client,” she says. When parents hand you their child’s future for the next several years, plus a hefty service fee, what they rely on has never been a polished mass email. It is meeting in person, talking, and the trust built up bit by bit.
She also tried acquiring clients through social media. She has done client operations and account operations, run a study-abroad account for the company, and also a UK food and lifestyle account. After going through the whole loop, she found that traffic does not equal client acquisition.
Xiaohongshu revamped its platform a while back. Study-abroad service content gets basically no reach without paid promotion, even if you used to be a big influencer. And if you do pay for promotion, the client acquisition cost is not worth it. The UK food and lifestyle account did have traffic, but that crowd was not the target audience and would not convert.
After going around in a circle, the two things that actually work are still: holding events and adding people on WeChat on the spot to leave a lead; and referrals from regular clients, where good service makes old clients bring in new ones on their own. AI cannot help with either of these.
AI does a different kind of work.
The one that left the deepest impression on Coco was connecting to the company database to automatically generate a beautiful admission poster. It can also produce a sample for a new student who comes in to consult, so they can see at a glance roughly what background it takes to get into a school at this tier.
But all of this is about boosting efficiency, not generating revenue.
“The company only pays for what helps us sign one more client. Something that just makes us work faster, it sees no need for.”
Keep the database in good shape, and employees will use AI on their own with the same result. So why buy a separate system.
Some people also ask whether AI can help write the application essays. That falls under academic misconduct, and large study-abroad agencies have partnerships with the schools, so they do not touch it.
I asked Coco what the moat of a large study-abroad agency is.
First, the brand. Big agencies will not abscond with the money like a small workshop might, so parents feel safe handing over their cash.
Second, domestic resources. They can take in clients brought in from within China, and can also serve younger students studying abroad.
Third, full-chain service. From the admission application, to after-enrollment tutoring, to future internships and job hunting, even for niche majors, they have a complete mentor database to catch the demand. A small agency can hardly provide this whole chain.
AI has flattened a standardized skill like writing essays, and in doing so it makes the non-standard things, trust and full-chain service, look even more valuable.
Standardized skills are not entirely without revenue potential either.
In the UK there are many one-on-one small classes, students are willing to pay, and customization demand is high, so there is now demand for a dedicated AI tutor to help answer questions during these small classes.
Some also want to use AI to help maintain relationships, reminding a consultant when a certain student’s birthday is coming up, so the relationship does not get cut off right after signing.
But she herself knows clearly that none of this is as direct as acquiring clients.
Beyond the pressure of performance, Coco has been feeling a bit lost lately.
One reason is that the whole industry is heading downhill. In those years when the economy was on the rise, a million-yuan annual income for a good salesperson was no myth. Now it is nowhere near that figure. AI has also made writing essays too easy, so proactive students simply handle the applications themselves and do not bother with an agency at all.
The other reason, more common for anyone working in the UK, is the work visa.
She once had a chance. She got an offer to switch jobs to a university’s international student recruitment role, which would be a lot easier; plenty of people in the industry move from the agency side to the institution side this way. The problem is that this kind of role does not sponsor a work visa. The role recruiting her was in Beijing, but her UK work visa had already been in the queue for years, the sunk cost was too high, so in the end she did not go.
Now, the queue time for a UK work visa is going from five years to ten.
“I do not even know if the decision I made back then was right.”
There are more and more things AI can flatten.
But what truly decides where a person ends up is still the things it cannot reach: who trusts you, who you trust, and whether you get to stay.
We always tend to romanticize the road we did not take.