Five ways to boost customer service capacity using AI ahead of peak
For retailers to make the most of the upcoming peak period, they have to not only acquire new customers but retain them. One of the key pillars that retailers need to get right is a great customer service experience.
Peak periods place extra demand on customer service. When we surveyed CS leaders we found that most (56%) would look to increase the capacity of their team by over 50% during this period. Hiring temporary staff is not only expensive, but can put your customer satisfaction levels at risk, impacting long-term revenue.
Embracing AI to help your existing team is an alternative strategy that retailers can take. Here’s how retailers are currently using AI to extend their customer service capacity.
1. Detecting ticket intents
One key aspect to resource planning is to know where you need to devote resources. Understanding what customer service enquiries (tickets) you get in a normal period vs. a peak period is essential to planning. Categorising tickets to analyse them can be done manually, but it is much faster, and more accurate, with AI.
AI can detect the intent in a query based on the keywords and where they sit in a sentence. For cycling apparel brand Le Col, using intent detection “helped us see really easily what the main bulk of our customer contacts are about… It’s given us a clear path to understand our priorities”.
From this point, retailers can then prioritise tickets when they come in, and assign resources more effectively.
2. Automatically resolve WISMO tickets
Once a retailer has categorised their tickets, more likely than not, WISMO (Where is My Order) queries are somewhere in the top 5. During peak, this number tends to go up disproportionately because carriers have more delays, warehouses are straining under extra capacity and so on. This then places an extra burden on customer service.
If AI is connected to your carrier and warehouse systems, it can easily look this information up for a customer and then communicate it back. Going one step further, a conversational AI can then confirm if they’ve received the order that’s been marked as “delivered” and if not, issue a replacement or a refund – all without a human intervention.
Skullcandy handle up to 80% of their WISMO queries automatically using AI, which allows their agents to focus more on troubleshooting and technical issues.
3. Use generative AI to answer FAQs
One of the major breakthroughs in AI is generative AI. This is the technology underlying Open AI’s ChatGPT, where responses from the AI are becoming more and more humanlike.
For retailers, one use of this technology is to train it on your FAQs and product information, so you can relay this to a customer in a conversational style.
Imagine - a customer is on a product details page and wants to know what material the dress is made from, or whether it fits true to size. A little chat window pops up, and the customer can ask that question, and get a coherent and true answer immediately. This removes an obstacle to purchase, and means that they are not waiting for an agent to come online and answer.
4. Improve agent processes
As well as answering questions for customers, AI can make the agent’s life easier by automating time-consuming tasks. The running brand On found that it took an agent 10-15 minutes to create a return label for a customer. By connecting all the disparate systems together, and letting AI effectively copy and paste the relevant details, this can be done in seconds, saving the customer service team 800+ hours a month.
5. Prevent tickets from occurring in the first place
Anyone stuck with a mountain of customer service requests knows that having multiple tickets from the same user can be draining. It means either the team have not responded fast enough and the customer is annoyed, or that something has gone wrong.
Waterdrop found that if a customer wanted to amend or cancel an order, it was luck of the draw whether an agent would see the request in time to action it. So they looked to automate this process using AI, meaning that the request got communicated as quickly as possible. This prevents a further 3-4 tickets being created when a customer receives an order they don’t want and wants a refund, a replacement or something else. According to Kane Sakata from waterdrop this: “saves our agents so much time in the long run, while making the customer experience smoother.”
Getting AI in before peak
If it feels too late to get started with AI before peak period starts, then don’t worry. Heritage suitcase brand Antler went live with a new team and new systems just before the peak period, and within a matter of weeks saw over 20% of tickets being handled by AI.
The key is to find a supplier that specialises in retail in order to get started quickly so you have templates and use cases ready to go. This will help you to take some of the pressure off your team and ensure your customers get faster responses.
About the author: Iain Moss is Head of Demand Generation at DigitalGenius.
In proud partnership with