Speech analytics is an emerging new solution that can improve both productivity and customer experiences in your call centre. While adoption is still low (11%), 15% of organisations say it is top or a key priority. 38% would like to use it for quality management, and 29% for compliance alerts. One of the quick wins that you can achieve using call centre speech analytics is keyword detection and analysis to arrive at customer intent or troubleshoot performance issues.
2 Types of Keyword Analysis Using Speech Analytics
Most speech analytics solutions support two types of keyword spotting methodologies – automated trend identification and targeted keyword detection. The former scans call recordings and transcripts to identify the most frequently spoken keywords and phrases. These are summarised in a report, which tells you what customers are talking about, their most common expectations and issues, and any dominant product trends.
The second methodology – the one under discussion in this article – is targeted keyword identification. You can configure the speech analytics solution to pick up on specific keywords based on your use case, market environment, and business objectives. Essentially, it lets you analyse an entire call without listening to it.
Top Keywords to Flag and Why
- Names of competitors – Mentions of a competitor might indicate that your product/service is lagging in a certain feature; repeated mention of a specific name could suggest the growing market dominance of one competitor
- Names of products/services – Product/service names help to segment call recordings into buckets and further analyse performance, opportunities, and challenges in that specific product or service category
- References to management – Statements like “I want to talk to your supervisor” are a clear indicator of dissatisfaction. It is vital to monitor this keyword trend to identify below-average performance and intervene with training
- Action words – Words like “want,” “buy,” “purchase,” “interest,” etc. signal cross-selling and upselling opportunities. The customer is happy with the product/service so far and would like to extend the brand relationship
- Obscenity or profanity – This is a must-flag keyword in any call centre, ensuring that agents do not breach basic quality assurance standards. Obscenity or profanity should ideally be flagged in real-time (for both the agent and the customer) so a supervisor can intervene
- Negative phrases – Words like “never” or “close” typically indicate a degree of dissatisfaction among customers. While it may not be the agent’s fault specifically, this keyword trend reveals the need to take broader action like reconfiguring your supply chain or improving the product
- “You have” – The phrase “you have” almost always refers to the performance of the agent who is taking the call. You can garner live and relevant customer feedback by flagging these instances and documenting what the customer had to say – like “you have been very helpful” or “you have not told me this”
Apart from this, make sure to run automated analysis as well so that you do not miss out on any emerging or trending keyword during call centre conversations.
