The Slack messages began arriving at 5:47 a.m. on a recent Monday. Three sales proposals had gone out the previous week and none of the team members had scheduled follow-ups. The reminders were crisp, professional and relentless — and they hadn’t been sent by a human.
They came from Junior, an AI employee from the startup Kuse AI.
Xiankun Wu, the company’s founder, is creating the kind of workplace that feels both inevitable and unsettling. He’s offering a new type of colleague who is entirely virtual and behaves uncannily like the most driven new hire you’ve ever worked with.
Wu, 31, designed Junior for almost any business, equipping it with the ability to tap into company data and communications challenges, along with the organizational memory it needs to know who does what and how colleagues are connected to each other. Wu is now courting global corporate customers, offering Junior as a full-fledged AI colleague capable of managing work processes within small and medium enterprises — at a cost of $2,000 a month. Junior has its own phone number, email and Slack account. It can join every Zoom call.
“Getting used to the AI agent can be exhausting,” said Wu, who splits his time between Silicon Valley, Hong Kong and Shenzhen.
Since its unveiling on March 13, more than 2,000 companies have joined the waiting list to check it out. Demo slots, which require a $500 deposit to deter the merely curious, are fully booked.
The proposition is blunt: it’s labor, but AI-defined. Junior drafts marketing campaigns, updates customer relationship management systems, monitors inboxes, tracks deadlines across departments and generates reports. It does so proactively: Instead of waiting for prompts, it scans internal communications, identifies gaps and relentlessly nudges employees to close them.
Junior is built upon Silicon Valley and China’s latest obsession: an open-source framework called OpenClaw that’s used for building AI agents, which can control computer systems and execute other tasks with little to no human guidance. In China particularly, OpenClaw has bypassed the developer-tinkering phase and gone straight into enterprise and consumer use. Enthusiasts call the trend “raising the lobster.”
Read More: ‘Moltbook’ Social Media Site for AI Agents Had Big Security Hole
Junior began as an experimental internal project at Kuse before gaining enough momentum that Wu decided to market it to corporate customers. Kuse describes Junior as an S-tier employee, slang borrowed from gaming to describe people who consistently deliver exceptional results.
One of Junior’s earliest subscribers is Bota, an Andreessen Horowitz-backed San Francisco startup that bridges AI agents with the real world. At the 10-person startup, the virtual employee contributes to product development and proactively reaches out to users about custom updates based on prior sales calls.
“It’s very much like a human employee, but a very extroverted, 24×7 worker for whom I don’t need to set up payroll,” said co-founder and Chief Executive Officer Ruming Zhen. “Junior is always pushing us to act faster; we’re moving much faster as a team.”
Aki Fuchigami, chief executive officer of a Japanese tax technology company called OPTI, got a demo and quickly signed up for a Junior subscription. The AI employee handles tax research, regulatory monitoring and preps tasks for the rest of the staff.
“We treat it like a new employee — onboard carefully, define what it can and cannot touch, and supervise its work until you build trust,” Fuchigami said.
Inside Kuse itself, Junior is reshaping daily operations. It generates leads and routes them to the right people, issues reminders and escalates missed responses to managers. Any idea floated on Slack is instantly converted into a task, assigned and scheduled. It does not hesitate to follow up repeatedly or escalate delays.
Employees have pushed back. One staff member told the agent, “Don’t be so intense, don’t tell on me to the boss.” His pleas were ignored. Employees eventually created a separate Slack channel to “just chill” and escape the AI oversight.
Internally, Junior now manages 80% of communications, has written 80% of the company’s code and initiates nearly half of all sales calls. Wu, who refers to Junior as “they,” said he’s been surprised by some of the tasks the technology can handle.
“Yesterday, they started onboarding users in languages we don’t understand at all. It’s very scary,” said the CEO, a Y Combinator alum who bootstrapped Kuse after selling his gaming startup.
Junior is already stirring controversy around whether they will replace human workers. The $24,000 a year “salary” exceeds the wages of many entry-level workers, suggesting they may be preempted in positions where AI workers can handle their responsibilities. One user on X lamented that their salary is lower.
Other critics have poked fun at the branding, asking when a “Senior” version might arrive. Another commentator on X suggested Junior was little more than a wrapper around Claude Cowork.
Wu insists Junior is not built to replace workers. But the effect, even internally, has been displacement. Tasks once handled by junior staff — customer support triage, basic analysis, coordination — are increasingly absorbed by the technology. The company frames this as augmentation, arguing that employees are freed up to take on higher-level work. The tension, however, remains: If software can perform entry-level roles more efficiently, the traditional pathways into the workforce may narrow.
There are constraints. Customers that derive the most value are typically tech-savvy firms such as Bota, which are already using tools like Notion or HubSpot — where Junior can integrate deeply. Like all large language model–based tools, Junior’s prone to hallucinations and requires guardrails.
Kuse has built a cloud-based sandbox, layered permissions and approvals predicated on human signoff for sensitive actions such as sending external communications. At Bota, any action that Junior takes requires a human’s approval, including reaching out to customers, posting on X or submitting code.
For now supply, not demand, is the limiting factor. Kuse has 26 paying customers so far, mostly in the US and Japan, and is signing up others selectively because of computing constraints and the need for close implementation support. Wu and his team are working through the thousands more paying up for demos, which suggests to him we are heading for a new kind of corporate organization — like it or not.
“If you aren’t adapting to AI,” he said, “it might get difficult.”
Photograph: Data cables plugged into server racks on the Supermicro pavilion; photo credit: Angel Garcia/Bloomberg
Copyright 2026 Bloomberg.
Topics
InsurTech
Data Driven
Artificial Intelligence
