About this case: The client is a cross-border logistics platform serving Hong Kong residents and expatriates living there. We haven’t named it.
Key Takeaways
- The platform’s customers contact support in the evening and at night. Its own team had gone home, and an automatic reply was the only answer.
- We staffed a dedicated team of office-based human agents on live chat and email, working in Chinese, English and Cantonese, around the clock.
- The team was trained on the platform’s own rules through a four-step loop before taking a live contact.
- In the trial phase, first-contact resolution rose from 60% to 85% and customer satisfaction from 72% to 94%.
At 10 p.m. a customer sees that her overseas parcel is marked as lost and contacts support. The platform’s team has gone home, and only an automatic AI reply comes back. She leaves several messages. No person answers, and she posts a complaint on social media.
Our account team uses that scenario to describe what the platform was dealing with. Its customers needed help at an hour when nobody was on shift.
A business whose customers write in at night
The platform handles parcel forwarding, bulk freight and cross-border returns. Its customers ask about declarations, tracking, problem parcels, pickup points and lost-parcel claims.
Three things were going wrong.
Nobody on shift at night. After the working day ended in China, customers in Europe and the United States were in their evening and active. Their requests piled up until morning, and lost-parcel disputes got slower and more expensive to settle.
Hiring in three languages. Hong Kong residents need service in Cantonese, and expatriates need it in English. Building that team locally meant high pay, office space, and repeated hiring when people left.
Slow answers to rule questions. How do I claim an undeclared parcel? What’s the fee for overdue storage? What happens with a mis-shipped item? Human agents who had to look each rule up kept customers waiting, and a wrong answer on customs can get a parcel held.
What we changed
Round-the-clock rota with a reinforced evening window
The team works a 24/7 rota. The 15:00 to midnight window, which the platform identified as its peak for European, American and Hong Kong evening contacts, gets extra seats.Bilingual hires, then Cantonese training
We hired people who already spoke Chinese and English and knew cross-border logistics. After they joined, they took Cantonese training for the platform’s service scenarios.A person behind every chat and email
Contacts on live chat and email reach a human agent working from one of our offices, including at night.One project contact for the client
The client deals with a single project contact. We handle recruiting, office space, attendance, quality checks and backfilling.How the team learned the rules
Coverage alone doesn’t raise first-contact resolution. A human agent who answers at midnight and says “I’ll check and get back to you” hasn’t resolved anything. So the setup included a training loop, run on our in-house training and assessment tool.
- One question becomes a hundred. For each high-frequency topic (claiming undeclared parcels, overdue storage rules, pickup-point storage, mis-shipped and missing items, lost-parcel claims), one original exam question generates around a hundred variants.
- Wrong answers bring more of the same. When a trainee gets a simulated question wrong, the tool keeps serving similar scenarios in English and Cantonese until the gap closes.
- The team’s weak points go to the project manager. The tool summarizes what the whole team gets wrong. The project manager runs a 15-minute review on those points every day, which is also how policy changes reach everyone at once.
- Two tests before going live. Training covers platform rules, scenario drills and cross-border knowledge. A trainee takes live contacts only after passing both the theory test and a simulated service session.
Research from other contact centers shows why this is worth the effort. SQM Group reports that each 1% improvement in first-contact resolution is associated with a 1% improvement in customer satisfaction. SQM doesn’t disclose the sample size, so treat it as a vendor research finding.
Results in the trial phase
Cross-border logistics platform: trial phase
The client’s management later wrote to thank one of the human agents on the team by name, noting how carefully she followed the platform’s requirements and how strong her performance figures were.
When this setup is more than you need
A 24/7 team with rule-level training is built for businesses where a wrong answer at night costs money, as it does with customs holds and lost-parcel claims. If your night contacts are mostly “where is my order” and your tracking page already answers that, a clear self-service page with a staffed morning queue may be enough.
If your night volume is a handful of contacts, a dedicated team will sit idle. Count contacts by hour for a month before deciding.
Three questions to ask about your own night shift
- At what local hour do your customers write in, and who is on shift then?
- When a night-shift human agent gets a rule question, do they answer it or log it for the morning?
- Before someone takes live contacts on your account, what do they have to pass?
Callnovo.ai staffed this account as a managed support team with multilingual human agents. For how we work with carriers and forwarders, see our logistics industry page. We cover language staffing in more detail in multilingual call center services.
Sources
- SQM Group, FCR Metric Operating Philosophy
- Callnovo, cross-border logistics support case (WeChat, in Chinese), September 17, 2026
Performance metrics reflect results from the trial phase of a Callnovo.ai partnership. Client identity withheld.