Reducing the team without compromising on the promised service means making those who stay work their vacation time.
Every summer, however, many teams do exactly that. The same deadlines. The same schedules. The same rules. With fewer people.
Everything runs smoothly until the first issue arises that falls outside the norm: an irregular refund, a held order, a customer who follows up. The person who can make the final decision is absent. The others have to improvise.
The point isn't to ask them to work faster. We need to decide before they start what can be automated, what takes priority, what requires human intervention, and what can wait.
Something to keep in mind: With fewer people, we promise less. But we deliver on what we do promise.
A smaller team inevitably changes the service. The problem begins when we have to wait for the first double charge before we admit it.
Start with the most common requests. How much time can you realistically spend responding to them? Who has the authority to approve a refund or a goodwill gesture? What happens if that person is absent? A question about tracking an order doesn’t require the same level of attention as a double charge or a delayed delivery. This discussion should take place before the holidays, not in the middle of an urgent request. It often leads to shortening business hours, announcing longer response times for low-risk issues, or appointing a substitute to handle certain decisions. Customers are much more likely to accept a slower but clear response than silence following an unfulfillable promise.
Our article on queue management is based on this reality: when capacity decreases, priorities don’t magically stay the same. As for service-level commitments, they’re only valid if they accurately describe the service that’s actually available.
Sorting requests by subject keeps your inbox organized—that's all.
When there aren't enough people, the real question is a different one: What does the person who submits the request do next? Four answers should be enough for them.
| Decision | When to take it | Requirement to be met |
|---|---|---|
| Automate | The question comes up often, and the answer remains the same | Switch to a human representative as soon as any information is missing |
| Handle first | Waiting can hold things up or be costly for the customer | A simple instruction that the whole team understands |
| Forward | A decision, an exception, or a sensitive situation requires a human | A designated representative and a complete file |
| Keep someone waiting | There is neither an emergency nor an immediate risk | A stated timeframe and a clear date for resuming operations |
The pitfall is deciding on a case-by-case basis depending on who’s handling the request. You end up with three different ways of handling the same issue and a team that’s constantly asking for confirmation. Write the rules down on a single page. If they require a twenty-page manual, they won’t survive the first week of August.
The scenario has only one test: At 4:00 p.m., standing alone in front of the line, does the person on duty know how to handle the situation, call someone, or announce a delay?
Summer doesn't make risky automation any more acceptable. It just makes it take longer to correct the mistake.
A request can be automated when the response is consistent, the data is reliable, and the action is authorized. As soon as information is missing, two sources contradict each other, or a business decision deviates from the rules, the tool must stop. It can prepare the file. It must not make up the decision. That is Klark’s role in a lean team. Klark can handle repetitive requests with a clear framework, gather relevant information, and hand off the case along with the existing history. It does not replace the absent person or their decision-making authority.
The boundary is determined on a case-by-case basis.
An order's status, retrieved from a reliable source, can be updated automatically; an exceptional refund will require human intervention. And if two systems conflict, Klark prepares the file and then stops.
Just because it's vacation time doesn't mean you can guess. Our article on artificial intelligence in customer service explains these limitations in detail.
At 10 a.m., a customer asks where her package is.
Five minutes later, another person writes: “I just got charged twice and I can’t pay for my order.” The second case is handled first.
Seniority matters, but it isn't the deciding factor.
Just look at what waiting can lead to: a customer stuck, a problem getting worse, an irreversible action, or a decision that no one on duty is authorized to make. A checklist with twenty criteria won't help either.
Every morning, the team starts with the high-priority cases, requests with no assigned handler, and those that have exceeded the stated deadline. The dual-track system skips over messages asking, “Where is my package?”—there’s no need to go through the entire queue to see them.
I'd rather have three priority levels that everyone uses than a perfect matrix that gets abandoned after two days. During the holidays, simplicity doesn't mean a lack of ambition. It's what keeps the rules from falling apart.
"Can you take over Léa's case while I'm on vacation?"
The message can be summed up in a single line; the investigation left to the replacement, however, is much more complex. This is not a handover.
The substitute must identify the problem in one sentence, the facts that have already been verified, the next step, and the limits of what they can decide.
If authorization is missing, the name of the person to contact must be included, along with their contact information. Is no one available to respond before the return? Let the customer know right away, instead of letting their request go in circles for days.
Also, keep a useful record of the conversation. Asking the client to repeat everything because the right colleague has left adds work at the worst possible time. A well-documented handoff conveys what has been understood, what has already been tried, and the expected decision. The test comes down to one question: Can a colleague who hasn’t followed the discussions take action within two minutes, without calling the person who is on leave? If the answer is no, the handoff isn’t ready.
At the start of the school year, seeing a zero on the counter is a welcome sight. The entire team can then clear the backlog in order, without wondering why certain requests lingered until the start of the school year or why the same questions kept coming up.
The scoreboard will be clean. As for the organization, it won't have learned a thing.
Start by cleaning things up. Some requests have been resolved in other ways. Others are no longer relevant. A few have been reopened because the initial response didn’t go far enough. And when the same question comes up twenty times, it may indicate information that can’t be found, a flawed procedure, or a good candidate for automation. The summer backlog clearly shows where the organization still relies on a single person. It also reveals rules that are too conservative—those that unnecessarily routed simple requests to human agents—and rules that are too broad, which allowed questionable cases to slip through.
Keep what’s worth keeping: better-defined responsibilities, a corrected information source, a clearer handover process, or a new scope of automation. On the other hand, stopgap measures designed to tide us over through August should not become the new default way of doing things.
Customer service teams know how to step up. That’s not the issue. The real test lies in a specific case: if the expert is on leave, does someone else know what to do without having to call them back? If so, the time off remains time off. Otherwise, the team on duty pays the price in exhaustion, and the customer pays the price by being kept waiting. Klark helps handle the repetitive tasks without changing your customer service tools, then passes the context on to a human as soon as a decision is required. It’s less spectacular than the promise of full automation. But above all, it’s much more useful when several experts are out of the office.
Your summer schedule is set, but is your customer service team still operating as if everyone were on the job? Let's talk about which requests to automate, which priorities to set, and which handoffs to prepare before the next round of vacations.
