Whether they are reaching out to the contact centre by phone or via chat, customers expect a more personalised and responsive experience than ever before. With the increasing number of customer interactions, how can companies reconcile the speed of processing with personalisation? What if semantic analysis was the key? How can the benefits be measured?
What is semantic analysis and how does it work?
Semantic analysis vs. syntactic analysis
Semantic analysis is a technique for analysing the 'meaning' of a text. The word "meaning" is important because the semantic analysis does not analyse the grammatical structure, as would the syntactic analysis technique, but rather the underlying intention of a sentence. This intention can also be characterised by a feeling or an emotion.
We could say that syntactic analysis is concerned with the "form" of a sentence, while semantic analysis is concerned with the "substance", i.e. the message.
So you think that the machine can’t detect all the subtleties of human language? Think again! Even if it is still impossible for the machine to interpret sarcasm, for example, significant progress has been made in this area in recent years. So much so that you are already using this technology (Natural Language Processing or NLP) daily through your favourite translation or word processing solutions.
Source: 8 Business Examples of Sentiment Analysis in Action (repustate.com)
In this example, an automatic process asynchronously detects messages based on certain keywords in social network threads. The semantic analysis then extracts the topic and measures the overall sentiment. After this step, the company has key information to adapt its treatment and customer relationship strategies to meet its objectives - reputation, resolution, and customer satisfaction for example.
Why is it interesting for your customer relationship?
The use of semantic analysis in your customer relations and your contact centre has many short and long-term advantages, especially if you manage all channels. Here are two examples: one of improving customer relations and one of improving employee relations.
- Improve your customer support by detecting self-service opportunities.
Semantic analysis can extract valuable information from unstructured data and is an advantage when deploying a self-service strategy. Self-service aims to handle certain customer requests autonomously and in real-time, thus relieving advisors of the burden of redirecting their time to more value-added processing. The analysis of the recurrence of certain questions will facilitate the implementation and use of dynamic chatbots and FAQs that are effective and relevant for customers.
Another application is detecting emails that have the same request and creating and proposing standard answers automatically. This will have a positive impact on the response time.
- Tailor advisor training
When you can break down text interaction by topic and issue, your agents' performance becomes quite obvious. You can quickly and easily identify problems and areas for improvement, and then create training and coaching for those specific issues. This means you can focus your efforts on training that create more value for your customers and employees.
What are the impacts on call centre performance?
Matching demands to available resources is a central issue for all contact centre managers. The objective is to minimise costs (number of agents on hold, processing time, transfer rate, etc.) while maximising customer satisfaction (net promoter score, attrition rate, etc.).
As you can see, it is through these strategic indicators that we can establish the performance of a contact centre and it is also through these indicators that we can measure improvements. With this in mind, what impact does semantic analysis have on these key indicators? Let's take a look at some of the key indicators for customer relationship centres :
- Reduce the transfer rate
Thanks to a better pre-qualification, incoming requests (such as emails for example) will be routed to the most qualified resolution points at a given time, thus reducing transfers between several people/departments whose fields of competence are sometimes very partitioned.
- Improving First Contact Resolution (FCR) :
How well is my contact centre performing? You can probably tell by looking at this indicator. Indeed, the implementation of a semantic analysis solution adapted to your activity should mechanically improve this indicator. According to our internal sources, 30% of customer interactions on average require an escalation or a call-back to the customer in contact centres (all sectors included) that are not equipped with semantic analysis. What about you? Do you know your FCR?
- Understanding the attrition rate
Attrition is a term used in many industries to describe the loss of customers over a period of time. The semantic analysis will not address attrition directly but rather will help the company to identify and understand the reasons for the departures through the last interactions that former customers had with the company in question.
- Increase Customer Satisfaction (NPS) :
The added value of implementing semantic analysis is multiple.
Firstly, this tool provides a finer granularity of information on your activity thanks to understanding verbatims’ meaning. A better understanding of consumer expectations and demands thanks to word clouds and also thanks to the analysis of emotions allows you to identify areas for improvement more precisely.
Secondly, the implementation of corrective action plans is facilitated by semantic analysis, is more easily measurable and results in better customer satisfaction.
Ready for implementation?
If you are convinced by these technologies to improve your customer relations, you will quickly have to choose among the many players in the field. How do you choose the one that will best support you? By taking into account the expertise and similar projects carried out, the constraints can be numerous.
Here are some of them:
- Data quality is very important in the implementation of these projects because the better the data quality, the more confidence users will have in the results they produce.
- The prior annotation of verbatims by business teams with a detailed knowledge of the business is crucial
- There are also numerous regulatory constraints on data anonymisation (local regulatory services and RGPD)
If you would like to find out how Worldline has implemented these technologies at La Banque Postale, click here.
If you would like to know more about Worldline's semantic analysis methodology, contact us now.
Discover also our Trusted Interactions solution to understand more how semantic analysis is working.