When the queue grows, the reflex is to hire. Often that’s not necessary. In most teams, a large part of the time doesn’t go to answering questions, but to everything around them: searching, conferring, distributing and duplicate work.

First measure where the time goes

Response time is an outcome, not a cause. As long as you don’t know where the minutes go, every measure is a gamble. So first look at three things:

• Waiting time before the first action. How long does a message sit open before someone picks it up?
• Handling time. How long does the answering itself take?
• Number of back-and-forths. How many messages does it take to complete one question?

Those three call for different solutions. Long waiting time is a distribution problem. Long handling time is an information problem. Lots of back and forth is a phrasing problem.

1. Let messages distribute themselves

Manually distributing the inbox costs time every morning and goes wrong as soon as someone is ill. Automatic classification and routing ensure a message lands directly with the right person or group.

The biggest effect isn’t in the time saved on distributing itself, but in what you prevent: messages sitting for hours because nobody felt like the owner.

2. Take the searching away

With many questions, most of the time doesn’t go to the answer, but to looking up the order, the earlier email and the agreement made on the phone. Three systems, three logins.

Make sure that information sits next to the message. A connection with your webshop or ERP often delivers more time savings than any text generator.

3. Let AI write the draft

Not to send, but to start with. A prepared reply that already includes the customer data, the order and the earlier conversations usually only needs to be checked and lightly adjusted.

Do make sure the AI draws on your own knowledge. Otherwise you get generic text you still have to rewrite — and then you’ve gained nothing.

4. Answer the question that hasn’t been asked yet

A lot of back and forth arises because the first answer is incomplete. If you reply to “where is my parcel” with just a track-and-trace link, you’ll get the follow-up question about the delivery day.

Look in your data at which follow-up questions come back most often and work the answer into the standard response right away. This is the cheapest win there is.

5. Prevent the question altogether

The fastest response is the response that isn’t needed. If order status is your biggest category, the problem isn’t your customer service but your communication before it: unclear expectations or updates that come too late.

From your contact reasons you can read exactly which questions you can remove at the source.

What not to do

Steering on response time alone. That’s easy to achieve with short, half-answered responses — and then the number of messages per question rises, while the customer is less satisfied. Always look at response time and at how many messages were needed.

A realistic expectation

Teams that go through this usually achieve a considerable reduction in response time, but the biggest part comes from points 2 and 4: less searching and less back and forth. The text generation is the cherry, not the cake.

Want to know where the time goes in your process? Book a conversation — we’ll look at your own figures with you.