Last update: 22 Jun, 2021
The use of AI in customer service does not per se mean the use of a bot. Nor does it mean that customers have as little contact with the employee as possible. It is about employees taking care of the demanding, complex requests of their customers, supported by AI and optimising the customer's service experience.
Questions about an order, changes of address or payment method, the current account or meter reading - standard enquiries of this kind reach the customer service of companies probably tens of thousands of times every day. Experience shows that standard enquiries account for 50 to 60 per cent of all enquiries. Processing them takes time, but rarely overtaxes the mind. Ideal conditions for using an AI-based bot to relieve customer service. In doing so, they ask themselves questions in advance such as "Is the investment worth it for our customer dialogues?", "How do I find the best AI strategy?" or "How costly is the implementation of a bot?"
Then come questions about implementation, for example, or about how much optimisation potential and time savings a bot actually brings in practice. This can be estimated quite well based on the type and number of requests and converted into paid working hours. Based on existing project data, it can be proven, for example, that a fully automated, case-closing transaction processing of an incoming email of medium complexity brings an average time saving of 20 minutes. So is fully automated customer service the solution? No. Because, sorry, but bots are not all-rounders!
Bots can easily work through purely repetitive tasks around the clock, 365 days a year - without holidays, without sleep, without quitting time and without a lack of motivation. But that is not the only thing that distinguishes them from employees. Everything that distinguishes capable employees is missing from them: Commitment, intellect, tact, the will to solve problems. So bots are not employees and, conversely, employees are not bots and should not work that way. They have other skills. AI cannot replace them. But it can assist, complement and support them. Even in the processing of complex tasks, such as complaints or technical support, there are subtasks that can be completed faster with AI than on foot by the employee - for example, the provision of information. Seen from this perspective, smart assistants are a practical tool, a tool for dealing with complex requests and workflows.
AI-based assistance solutions are able, for example, to recognise the content of a telephone enquiry via speech analysis and immediately take down relevant content. This saves the employee having to take notes - and also the mistakes he or she makes when taking notes, such as transposed numbers or the like. Instead, they can concentrate fully on the telephone conversation. The AI assistant searches the CRM for the correct data record and displays it immediately in the client. And the smart assistant not only transcribes the spoken word, but also recognises relevant information, which it enters directly into forms, for example.
If relevant content of the conversation is recognised, such as a product number, an invoice number or keywords, a smart assistant can immediately display the corresponding or matching content and thus immediately "put the employee in the picture" about what the customer is concerned with. Instead of annoying searches in different systems, all important information is immediately available to the employee. This brings momentum into the daily service work and also clears the way for cross-selling and upselling. And with the right talent, this is certainly more fun than changing a hundred addresses or typing in meter readings for days on end.
And there is even more that a bot can NOT do: Be happy about a successful sale, a won back customer and a mutually positive, friendly customer contact - and be motivated for the next one!
Author: Susanne Feldt