Some people may fear that artificial intelligence is becoming increasingly dominant and will start taking away jobs, leaving people unemployed. What we can focus on is looking at which jobs AI could actually do better. This would leave more time for humans to focus on their strengths. This includes giving way to automated call topics.
In terms of customer service and call centres, there are a lot of agents all across the globe. They are manually collecting data and making notes for call notes. Imagine if data entry would be automatic. The agents would focus on communication, counselling, and customer service. How much better could that make customer service? Not only would the customer service win. Getting exact and automated data would also mean more knowledge to back up business decisions.
In this post, we would like to shed light on the improvement possibilities that automated data collection could give. There is an impact on customer service, business decisions, and product owners.
Manual data entry and notes may be insufficient and not trustworthy compared to automated call topics
When a customer calls a call centre, the agents have to follow certain steps to document calls. There are many reasons behind that. It is needed for call separation into topics and making sure that the information is there for repeat calls. Another need is for gathering data for business analytics.
Each company has its own rules that they ask agents to follow when it comes to documentation and making notes. Some companies only ask for the agents to mark down the topic of the call and resolution. Others require a lot more documentation.
Even if the requirement is to write down the reason, actions and resolution, that may not all end up in the note. Some agents tend to write short notes, others will write essays. Gathering data for analysis based on such notes is pretty difficult. The content and amount of detail will be subjective, so there should be a better way. In the view of the agent, the more that they are asked to do, the more difficult their job is.
From a personal perspective, I have worked as a customer service agent for a few different international companies. The documentation rules varied, but it was always clear that the topic had to be documented together with a short description of the customer contact. Agents had to document and save the information in CRM before the customer contact had ended. This meant that documentation had to be simultaneous with the call. That made it much more difficult to keep the focus 100% on the customer. The time pressure was an issue for a lot of customer service agents. This left them with either poor documentation or lower call quality. I now understand how automated call topic generation for documentation could have made my life much easier.
How Feelingstream was born – from necessity
When conducting notes, some companies only request their customer service agents to define the reason or topic of customer contact. This is how Feelingstream actually was born as well. Feelingstream’s CEO Terje Ennomäe was working for Bigbank. She created a solution for registering customer contact reasons (read the story behind Feelingstream).
For example, in the consumer loan business, the customer contact themes or topics may include:
- General information about the loan (conditions, interest rate)
- I’m applying and I have additional questions (documents, deadlines)
- Waiting for a decision to get a loan
- I got a positive decision, signing the documents, but I have a question
- Why was my decision negative? I’m not sure I understood the response.
- I want to change my payment schedule (partial or full repayment)
- I’m in debt, please help me try to fix this
- The information displayed on my credit file is incorrect
When creating such a list of reasons for agents to choose from, the options come from an analysis of past contacts. Having a drop-down menu may seem great, but it has its limitations.
Manual vs automated classification
When Terje was first starting this contact classification project, she added an option “other” for the agents as well. The first report showed that this was a mistake, as agents categorized 80% of calls as that. This data was a clear indication that contact classification had to change. Instead of relying on customer service agents, Feelingstream has created a more objective, smart technical solution to do that instead.
Feelingstream transcribes calls into text, meaning that users can easily analyze and review classified calls. Data is gathered across multiple channels, so information from calls, emails, and chatbots can all come together for analysis. At first, our data scientists teach the AI solution. It needs the training to understand the patterns and get the classification going. With some work, the solution can be up and running quickly, providing the data fast and accurately. Automated call topics are just part of what the automated system can really achieve.
What is the real impact of using Feelingstream’s solution?
Automated call topics offer much more accurate classification of contacts. The classification data is more relevant and helpful for business decisions.
When calls are classified and there is a rise in calls on one topic – based on the previous example, let’s say topic 3 “Waiting for a decision to get a loan” – then it is clear that something should be done about it. It could be something that makes the process more efficient so that the customers get their replies faster. Another possible option would be to keep the customers updated and send them emails. Emails can advise how far along their loan process is and when they should be expecting a reply. Maybe it is some third or fourth solution or a combination that makes for the best business. All in all, knowing the reasons behind the calls and acting based on that data makes room for thought-out changes and business decisions.
Another example – data shows that there has been a decline in calls on topic 5 “Why was my decision negative? I’m not sure I understood the response”. Such a decline in contacts shows that the automated communication explaining the negative decision has been sufficient. The team can be proud of the clarity of the process.
From an agent’s perspective – if you take off the pressure to multitask, their work gets easier. The agents can focus on communication, active listening, and providing excellent customer service. If the agents know that the documentation is automatic, they can just actively listen and resolve customers’ issues. This in itself can make them greater at their work.
Call length can decrease up to 25% – time won from not having to document calls manually. With automated call topics, Agents can take more calls and be more focused during those calls.
Process leaders and product owners can get more exact data and therefore can make their solutions more effective. For example, take topic 1 from our example list “General information about the loan (conditions, interest rate)”. Analyzing what the customers ask about will let the product owner know where to add information and what to clarify. When all of the information is available on the website, in all advertisements, campaigns, etc., the likelihood of the customer call is minimal.
Bettering business processes based on actual needs and keeping up with the latest data, always thriving and working towards efficiency and clarity of processes.
Taking action based on understanding call reasons means that First Contact Resolution and customer satisfaction rates can go up.
Customer satisfaction rates could go up 10%.
Customer service agents that can focus on customers, will be more efficient and able to provide better service. Processes that are updated and created based on customer needs are more transparent and clear. Every automatized process is a win for the company. If you wish to achieve all of this with the help of Feelingstream’s AI solution, please sign up for a live demo.
Also, make sure to check out our Ultimate Guide to Efficiency with AI.