A financing brief from Jiemian News dated 2026-11-02 said RoboScience planned to use proceeds from the round for validation and deployment in settings including retail and industry. This is a stated use of funds, not evidence that deployment or staff savings have already been achieved. For technology companies moving into actual delivery, founders still need to explain who will handle the work left after a system takes on more tasks.

In its financing announcement dated 2026-10-29, Serval listed service desk requests and employee onboarding and offboarding as automation settings. This describes the company's intended applications. We have not tested its product and do not infer that automation has eliminated human work. The overseas material is for background only. As an FA adviser, Youdu Capital serves domestic financing for Chinese technology companies and does not provide cross-border financial services.

The discussion below concerns a conditional operating situation, not the staffing of either company. Suppose a system first handles tasks with clear rules and complete information. The number of cases requiring a person may decline. But cutting staff coverage in direct proportion to that decline could turn a real efficiency gain into a new delivery bottleneck. The remaining tasks no longer resemble the original mix.

The first causal step is task selection. Once a system takes work that can be clearly described and completed through established steps, people may receive more cases with conflicting information, unresolved conditions or a need for exceptional judgment. The composition of the human queue can change even without growth in overall business volume. An old average handling time included many simple tasks and cannot directly determine the hours required by the new queue.

The second effect concerns capability. Complex cases may require someone who knows the setting, can ask the right follow-up questions or can organize further action, rather than any available employee. If the company first reduces the hours during which that expertise is available, remaining cases may queue behind a few people. Fast handling of routine matters can coexist with long delays in exceptional ones. Fewer cases therefore need not reduce human workload by the same proportion.

Consider a fictional retail replenishment process. A system completes standard orders, while people address substitutions for unavailable stock, conflicting information and special customer requests. Employees previously fitted exceptions between ordinary orders. After automation, fewer exceptions might nevertheless fill the available hours of experienced staff. This example illustrates a change in task mix. It does not suggest that either company faces such a problem or compare product performance.

Before changing staff coverage, founders can observe the handoff queue separately: why each task was transferred, how much actual handling time it required, how much time it spent waiting and what experience was needed to proceed. Records should come from normal operations, preferably using categories and de-identified information without copying additional sensitive customer material. Waiting for further information should be separated from active judgment, so external delays are not all mistaken for a staffing shortage.

Observation must also include work that nobody has picked up promptly. If reporting covers only completed handoffs, the hardest cases may remain in the queue while average handling time appears to shorten. Founders can examine why unfinished cases remain and what subsequently happens to them, distinguishing work that is no longer needed from work that has merely disappeared from completion statistics. The purpose is to identify remaining work, not prescribe one time limit for every queue.

Staffing discussions should also focus on that remaining work. Experienced employees can have explicit periods reserved for handoffs, support roles can organize information, and the reasoning used for common exceptions can gradually be documented. Naming a contact does not establish that the person is always available. Whether a new team member can handle cases independently should be checked through specific tasks; finishing training alone should not be counted in advance as full usable capability.

This does not require a company to retain its original staffing forever. If observation shows that remaining tasks are also clear and their burden keeps declining, the organization can discuss further adjustments. What matters is keeping savings on routine tasks separate from capability for complex work. The former indicates how much repetition can be removed; the latter determines when remaining matters can be finished. Each needs its own explainable basis.

Over the next three to six months, a company planning wider automation can begin with one process and retain handoff categories, actual handling hours and unfinished-case records before and after expansion. If human case counts fall but waits for experienced staff increase, first examine whether remaining tasks are harder. Then decide whether to improve transfer information, provide more coverage for handoffs or narrow the expansion, rather than immediately realizing planned savings from case counts alone.

If the handoff burden does decline, confirm that cases have been properly completed, rather than difficult requests being deferred, customers being steered away or employees finishing work outside scheduled hours. Adjust staffing gradually and continue observing whether new settings change the task mix. This conditional outlook sets no universal success ratio. New automation and the capability retained to handle the remainder need to be open to review together.

Financing materials can explain separately which tasks the system handles, which situations still require human judgment and the basis used to estimate personnel costs. As an FA adviser to unlisted companies, Youdu can help organize domestic financing materials, without replacing technical evaluation or investors' judgment. This is general operating-method analysis. It promises no efficiency, delivery or financing outcome and is not specific investment, legal or tax advice. Consult appropriately qualified professionals on customer agreements and personal information handling.

Sources 界面新闻:机器科学融资与场景验证计划(东方财富转载) (2026-11-02) https://finance.eastmoney.com/a/202611023918484738.html Serval: We raised $150M to make everyone a builder. (2026-10-29) https://www.serval.com/blog/we-raised-150m-to-make-everyone-a-builder

Content note: This article was drafted with AI assistance and reviewed through an independent AI process, using the listed public sources. It contains no original interviews, and no first-hand testing of the products, services or methods discussed was conducted. Checking public sources is not a field investigation or first-hand experience.