UK water utilities face potential fines of up to 10% their annual turnover if their customer experience strategies fall flat, warns Phil Smith, CEO at QPC Group.
The water industry continues to struggle to meet the performance commitments it made to customers and third-party service providers more than five years ago in PR19.
Findings from the C-MeX (Customer Measure of Experience) and D-MeX (Developer Measure of Experience) results for 2024-25 show that less than half were awarded a financial payment for meeting or exceeding their targets by Ofwat.
However, even these results may not give a true picture of customer service levels because C-MeX is based upon random samples carried out by a market research firm and so does not capture the entirety of the customer experience.
It’s a problem that’s also experienced internally, with many providers relying on Net Promoter Scores (NPS) or Customer Satisfaction Surveys (CSAT) to determine whether their customer service is delivering.
While these do provide some insight, the scores don’t contain any data about the customer journey, so offer limited value when it comes to diagnosing issues and improving processes. It’s for these reasons that providers need to look to capture the entire customer experience but this presents challenges.
Missing data
To start with, most customer channels, be it chat, email, IVR, or phone, tend to be handled as distinct channels and use separate platforms, resulting in siloed data.
What’s more, because these platforms tend to capture complete session details they miss failed contacts, multiple customer calls, partial resolutions, escalations, failed deflections and all the friction between systems and channels.
So, while the customer may perceive the company as offering an omnichannel experience, these processes are in fact splintered, leading to partial or incomplete data sets and poor customer insight.
In contrast, if the provider can capture all of that data, no matter how inconsequential, together with additional system data related to case management and field operations teams, it becomes possible to gain a true understanding of where processes are working and where they are failing and causing friction.
If this data can be accrued and analysed in real-time, those insights can provide the agent with contextual data and sentiment analysis so they are informed and prepared so can deal with the enquiry more effectively.
Frustration as a metric
To do this the provider needs an overlying agnostic architecture that can integrate with these disparate systems in real-time. Contacts can then be intelligently routed and escalated with all of the waypoints from the customer journey while analysis can be used to determine sentiment.
For example, it’s possible to create a Customer Effort Score (CES) which reflects the amount of friction the customer has encountered along the way.
Each event, be it ineffective interaction with IVR or a chat bot or being transferred or placed on hold, is awarded a weight by an algorithm according to the intensity of the effort and when it happened sequentially.
The final score represents this effort along with root cause and context for every point of contact, providing the agent with a complete record of the customer journey and Operations with the information they need to be able to address bottlenecks and refine the process.
One provider, who serves close to three million users in the UK, recently used this approach to carry out root cause analysis overlaid with total call handling time (TCHT) and CST data to improve its processes.
It was able to use this data in a myriad of ways, including identifying repeat contacts by category, sources of failure behind repeat contacts, contacts that could benefit from channel ‘shift’ or automation and those contacts responsible for high Average Handling Times (AHT).
As a direct result, the provider was able to support future planning models (such as billing) through its better understanding of contact cycles and went on to address skills gaps among its advisors by comparing AHT against the contact types being handled, leading to a 10% reduction in AHT.
Adding AI
Of course, there’s also now the role of AI to consider. AI chatbots need to be able to access this kind of data to be able to intuit intent, for instance, so they can hand over to an agent when it becomes necessary to do so.
Their responses will only ever be as good as the data they have access to and that’s going to be an even more pressing issue as we move to agentic AI, where the bot will need to independently retrieve and resolve issues. But while AI will play an important role in CX, it’s also necessary to consider how it will be monitored and its success measured. AI agents evolve over time and so can degrade in the quality of their response so the provider needs to be able to capture and analyse AI exchanges too. Should the persona of the bot begin to go off message, a human-in-the-loop (HITL) can then be used to correct and retrain the AI.
In order to oversee all of these processes effectively, it’s vital that water companies begin to look at how they can harness and use all the customer data at their disposal, particularly given the introduction of Condition G to their licences.
Announced last year and further updated in March, it requires providers to communicate proactively, be easy to contact, provide support, and learn from past experiences to improve delivery. Those that fail in this regard can potentially face fines of up to 10% of annual turnover, making a unified CX infrastructure a priority.
