13th of June 2024
Increasing Customer Satisfaction in the utilities sector
Last year, Ami partnered with a well-known water utility provider to improve their customer satisfaction rating. In a sector where customers are often reporting a problem with their water supply, it is notoriously challenging to secure a high CSAT score.
To combat this, we employed a variety of strategies, resulting in a much-improved customer experience and a current CSAT rating of 7/10. In this case study we will examine how Ami has increased customer satisfaction for this company.
Strategies to improve customer satisfaction
Actionable and concise wording
Let’s face it: if your stress levels are rising because you have no water for your morning shower, the last thing you want to do is talk to a chatbot. When many queries have a negative starting point like this, Ami needs to assist the customer as quickly and as effectively as possible.
Ami therefore sets out to self-serve as many customers as possible, guiding them to the correct information on the client’s website in a conversational manner. To accomplish this, Ami first establishes what the customer needs by asking relevant, constructive questions:
In this example, Ami empathises with the customer before immediately directing them to the live incident map. Ami then sets to work gathering the information required in order to provide the correct piece of final advice, tailored to each customer.
Our research has also shown that providers not having actionable content leads to customers dropping out of conversations early and not following through with the information provided. In contrast to the example above, wording such as ‘You can find news about your water supply on our outage page’ often struggles to engage the customer and therefore is less likely to lead to a positive outcome.
Combined with Ami’s high level of understanding, we have found that concise but actionable wording leads to a higher level of satisfaction.
Addressing missing content
In order to meet the evolving needs of customers, Ami must be adaptive. In addition to reading the information on the water utility provider’s website and with the client’s input, Ami is regularly updated with new content and knowledge, guided by recent customer queries and behaviour.
When this utilities provider came on board, our data suggested that up to 30% of queries fell into this missing content category. From our ongoing refinement work, our team has reduced this figure to below 10%. Addressing missing content also serves to prevent misunderstandings, meaning that the advice Ami offers is more likely to be helpful and will never provide incorrect advice.
This advantage is in stark contrast to most common chatbots which are constrained by buttons and therefore only allow the customer to choose from a handful of pre-set issues. Inevitably, this results in misunderstood queries and a very unhappy customer. Ami meanwhile holds over five hundred outcomes, a figure that is growing continuously in line with the queries that customers actually ask.
Retaining context
If you’re dealing with a leak or have no water, the last thing you want to do is repeat yourself. Ami prevents this frustration by retaining context from earlier in the conversation. As shown in the example below, Ami recognises that asking if this customer’s neighbours are experiencing the same issue would be redundant. This results in a more efficient and conversational exchange for the customer.
Triaging customers to the correct team
Inevitably, Ami cannot solve all queries. However, Ami should recognise when escalation is required, and correctly redirect the customer to the relevant team. For example, for outages that have not previously been reported, customers are instructed to ring a certain number. Vitally, Ami has the ability to separate urgent issues from standard ones, ensuring urgent queries are escalated or provided with the correct actionable advice.
When redirecting, Ami purposefully uses actionable language (e.g. ‘by clicking here’). This is informed by our previous research in this sector, telling us that these overt linguistic choices make it clearer for the customer to identify the action they need to take to get their problem sorted, whether that be clicking a link or phoning a number.
Considerations when measuring customer satisfaction
CSAT scores are not always a reflection of the quality of the conversation. Whilst we can employ strategies to increase this rating, our research has demonstrated that CSAT rating is also a reflection of other external factors.
Negative topic matter
Understandably, if customers are stressed about having no water, their patience levels will be running low and the matter at hand will be high priority for them. It’s also worth remembering that the advice Ami gives is set by the client and will not always be well-received. In other words, if the customer is not satisfied with the advice despite Ami providing the correct information, the CSAT rating is still likely to suffer.
Chatbots are not human
Most customers realise early on that Ami is not human. This can affect the mindset and even the behaviour of customers, with many holding pre-conceived negative opinions concerning whether a chatbot will be able to address their issue effectively. Some users will therefore be unwilling to engage with Ami, making Ami’s job more challenging.
Another consideration is that we are naturally less likely to feel the need to rate a chatbot highly – it's not like they have emotions, right? Conversely, we may assume a human agent would benefit more from a positive rating in terms of earning a bonus or gaining job satisfaction.
Unrated chats
More than 75% of chats are not provided with a CSAT rating. Our research shows that people tend to rate chats negatively when they have had a bad experience, but not highly when the experience is good. And even less so when talking to a chatbot!
With so many unrated chats, it’s not unreasonable to assume that some customers just did not feel strongly enough to provide a positive rating, even when Ami deserved it.
Unbiased data
Unlike our competitors, we never proactively ask customers for good ratings, meaning the CSAT scores that Ami receives are unbiased.
There’s also no hidden agenda, we don’t pick and choose when Ami reminds customers. Ami will ask customers for a rating at the end of every conversation, good or bad.
Summary
With the help of the strategies above and based on our research of previous installations in the same sector, we have estimated that we have improved CSAT for this water utility provider by roughly 200%. Ami’s average CSAT score is now 7/10, with over 40% of feedback awarding the highest rating of 5. Using Ami’s capabilities and the multitude of data available to us, our sights are firmly set on building on this achievement and raising CSAT even further for this client.
Author
Jessica Draper
Comments