Back to articles

How AI changes customer contact in small firms

AI•4 min read

A company with ten employees rarely has a customer service department. Customer contact is handled by whoever happens to be free: a salesperson, a receptionist, the owner. Emails are answered between two meetings and the phone is picked up when someone has time.

AI does not change the fact that customers have questions. It changes what happens while you cannot answer, and what you know when you finally do. Here are four concrete changes, without any big promises.

The first reply comes straight away

What a customer remembers is usually not how good the answer was, but how long it took. A question about delivery time that is answered after two days has already cost a sale, because the customer has had time to ask someone else.

An AI chat on the site answers while the visitor is still there. The answer is taken from what you have written yourselves, for example terms, services and common questions. That means the simple questions never have to wait for a person.

Evenings, weekends and lunch

A car dealer gets many of its enquiries on Sunday evening, when the customer is at home comparing. An accounting firm gets them in January at eleven at night. Without an answer they land in an inbox and wait until Monday.

An AI chat answers around the clock. Outside opening hours the site can take a message with a number and a matter instead of showing a queue, and you call back when you can. An AI receptionist can also take phone calls when you are busy, ask what you need to know and save the matter with a summary.

It is not about replacing someone who goes home at five. It is about the matter not getting lost when nobody is there.

The team gets better information

Someone who calls a customer without preparation starts by asking what it is about. That is a poor start, because the customer has already explained it once.

When the AI has made the first contact, the matter can follow along: what the customer asked, what was answered, contact details and a short summary. Whoever calls can read three lines and start in the right place. A dental practice then knows it is an urgent appointment and not a routine check-up, before the phone even rings.

People get time for what needs them

Many questions are repeats: opening hours, prices, how long something takes. Answering them is necessary but rarely rewarding for the person answering.

When that kind of question is answered by the AI, more time is left for the calls where a person makes a difference: a quote that needs explaining, a customer who is unsure, a problem that calls for judgement. That is where the staff's knowledge and tone matter.

It requires one rule: the AI must know when to hand over. If the answer is missing from the material, the customer should get a call, not a guess.

What you need in place

AI is only as good as its source material. What you should do first:

  • Write down the answers to the ten most common questions in one place.
  • Decide which questions should always go to a person.
  • Decide who calls back, and within what time.

Start small and read the conversations from the first weeks. They show what customers actually ask, which rarely matches what you thought.

How it fits together in Leadcaller

The Leadcaller widget brings this together on your site. The AI chat answers from your site and knowledge base, the visitor can be called back within 45 seconds or book a time, and after every call you receive contact details, a summary, a transcript and a recording by email and SMS. The AI receptionist, available in beta, takes the calls on the phone. The widget is installed with one line of code.

Key takeaways

  • The first reply comes straight away, including evenings and weekends.
  • The team gets the question, the answer and the contact details together before anyone calls.
  • People spend their time on the calls that need them.