Interview – Energy Utilities: How to Get from Service Desert to Beach Bar
However, according to a recent PwC study, “the energy sector is still at an early stage when it comes to 2.0 services.” At the same time, “the continuous advancement of self-service offerings for customers is a key industry driver among utilities.” Going forward, “AI-based assistance systems will increasingly play a supporting role.”
For more than 25 years, CreaLog in Munich has been developing exactly the kind of solutions that take digitalization and telephone self-service to a new level for companies across a wide range of industries. With AI-based VoiceBots – formerly known as voice dialog systems, voice portals, or Interactive Voice Response (IVR) – a new generation of systems is now in practical use at utility companies (EVU). These systems improve reachability and service quality for customers while also automating internal processes effectively. Trade journalist Nicolai Jereb spoke with CreaLog Key Account Manager Stefan Riesel about current projects.
Nicolai Jereb:
In which areas does CreaLog support digitalization in the energy industry?
Stefan Riesel:
We work in three core areas. First, of course, there's customer care in the broadest sense – including phone-based self-service through AI-powered voice systems. Customers can report their meter reading, change their address, adjust their advance payment, or get answers to FAQs.
Second, there's the important area of incident management for control centers and service hotlines, which come into play when grid outages affect many customers at once. Callers are pre-qualified, informed about outages, and if they wish, even called back automatically once the outage has been resolved.
The third area concerns internal processes in the utilities' grid control centers. This involves systems for recording and documenting calls related to load distribution, power trading, and the management of grid elements.
Potential for Utilities
Self-Services for Customers
Control Center & Customer Hotline
IT Hotline & Infrastructure
Report Meter Readings
Change Addresses
Change Bank Details
Adjust Advance Payments
FAQ
Top Up Prepaid Meters
Customer Satisfaction Survey
Pre-Qualification
Outage Announcement Management
Connection/Disconnection System for Grid Elements
Callback After Issue Resolution
Call Recording
Password Reset
Trouble Ticket Management
Alerting
Pre-Qualification
Call Routing
Queuing
Unified Messaging
Announcement Before Recording (Section 201 of the German Criminal Code)
Call Recording
Let's start at the beginning: why voice-based customer self-service at all?
There are several reasons for that: as digital voice assistants such as Siri, Alexa, or Google Home have become established in everyday private life, acceptance of voice control has grown enormously. Users are willing to communicate with machines and welcome the option when the service is fast and good.
But aren't there chatbots for that these days?
Sure, but for most customers, reaching for the phone comes a lot more naturally than typing a message on a PC or smartphone. And whatever a smart chatbot achieves as a text-based dialog system, an AI-powered VoiceBot can easily replicate over the phone. From practical experience, I can confirm that customers at many municipal utilities are happy to let our VoiceBots help them.
What exactly can these solutions do, and what else is possible?
At one utility, a completely new customer service system is about to launch. In the first project phase, we're implementing caller pre-qualification through natural-language dialog. This technology is called NLU, short for Natural Language Understanding, which roughly translates to natural speech recognition. So the VoiceBot understands the spoken word, however the caller happens to phrase it.
For simply entering a customer number and name, that might sound fairly basic, but the next possible project phase goes a step further: working closely with service staff, we can then teach the VoiceBot how callers phrase their requests across a wide range of topics, and what the best responses look like to resolve the request on the first call.
This allows meter readings to be submitted over the phone in a highly intelligent, self-service way, including a plausibility check and direct handover to SAP IS-U or other backend systems.
Our VoiceBots can also handle address changes, answer FAQs on topics such as moving house or smart meters, change advance payments, or prepare a contract switch.
Does this only work for standard topics?
Not only, no. But experts estimate that 70 to 80 percent of all service inquiries at utilities are exactly these simple, standard questions – ones that a VoiceBot can answer in a well-structured, automated way. Quickly, without hassle, around the clock, seven days a week.
That means the utility is always reachable, and no customer has to sit stuck in a phone queue or call back the next day. VoiceBots, after all, are never annoyed – they're consistently friendly and helpful.
Does that make human employees redundant, then?
No, not at all! Service staff are relieved of standard requests and get the chance to focus, without time pressure, on the topics that require real consultation – individual energy advice, for example. At the same time, this kind of up-to-date customer service also positions the utility as an innovative energy provider for a young, mobile, smart audience.
And what happens when the VoiceBot doesn't know how to proceed?
Good that you bring that up. Typically, the caller is then transferred to a personal contact, of course.
And the customer has to start explaining their request all over again?
No – that was unfortunately often true in the past, but only because the information from the VoiceBot wasn't being passed on to the service employee. Today, the so-called 'digital handshake' – the automated transfer of the information already captured – has to work flawlessly. Otherwise, the whole customer experience falls apart fast.
That's why, for most of our customers, our solutions are integrated into the utility's systems. This ensures call data is reliably passed on to the employee, and the customer can pick up their request exactly where they left off with the VoiceBot.
So can energy utilities just go all-in with a VoiceBot right away?
If only (laughs). Even for AI-based VoiceBots, thorough, careful upfront planning is essential. So we recommend starting with small steps to gather experience, and continuously asking customers how they're getting on with the VoiceBot and whether they're satisfied with the service.
The insights gained from that naturally need to be acted on right away – that's part of any sound quality management approach. Ultimately, it always comes down to acceptance of automated customer dialog.
And how does a utility recognize the right implementation partner?
The potential partner should have as much experience as possible in this field, avoid promising the moon, and offer fair, realistic proposals. On top of that, they should have verifiable references, and their staff should be genuine experts with the right know-how.
Let's turn to incident management, a critical area for every utility. Specifically, how can you help here?
Let's walk through the process step by step: first, the caller reaching the outage line is greeted by the voice system. If there's a general supply outage, they receive the relevant information right away. If not, they can optionally enter their postal code to be informed about any known local outages. If they're reporting a new outage, they're transferred directly to a human employee.
How does the system know about the outage, and who takes care of organizing the required announcements?
That's fairly straightforward: when an outage occurs within the grid infrastructure, a predefined alert chain is triggered. The central control center records the location and scope of the outage and passes the data directly to the voice system, either through our management suite or a direct interface.
So if, for example, a caller reports an interruption to their power or gas supply at home, the system checks the database to see whether the outage is already known for the postal code area they've given.
Our customer Netze BW, for instance, offers this service to all utilities in Baden-Württemberg – around the clock, seven days a week. This solution demonstrably handles up to 80 percent of outage reports fully on its own. That's genuinely impressive!
That leaves the third area…
Right, exactly. This one covers the utility's internal processes in the grid control centers – topics like load distribution, power trading, or switching calls, meaning the connection or disconnection of grid elements. Under Section 201 of the German Criminal Code, these calls must be recorded. Our solution greets the caller before recording starts and informs them that the entire call will be recorded.
That way, there's proof later, if needed, of the agreements and arrangements made. After all, this concerns the highly sensitive area of grid infrastructure – switching calls, meaning the disconnection and reconnection of lines, which touches on safety-relevant matters for the staff working on those lines, and in some cases simply involves large sums of money in buying and selling energy.
Broadly speaking, our solutions provide the technical foundation for a wide range of applications, such as pre-qualification, call routing, queuing, unified messaging, or customer surveys – including password resets for employees or management alerting.
What's your takeaway on digitalization at utilities?
A little tongue-in-cheek, you could say modern self-service offerings take utility customers straight from the service desert to the beach bar. Put more soberly: VoiceBots improve customer dialog, enhance the customer experience, and build stronger customer loyalty – all while cutting costs at the same time! Digitalizing internal processes in the control centers ensures transparent, auditable phone communication in the sensitive, capital-intensive areas of energy supply.
For our roughly 40 customers across Germany and the EU, this is already working very well!
