The rise of voice-controlled interface

It is perhaps not surprising that voice has become such an important interface. It’s as natural and as human as it comes. While voice has been around for a long time, its success has always hinged on the quality of the language processing, the problems we were seeking to solve with voice and a killer product.

As we know, that all changed with Amazon Alexa, Google Home, Siri and Cortana; the world of interaction has changed and now “we” can’t go back. Centrica has been at the forefront of voice adoption from the get-go – from voice-enabled IVR using voice, or piloting voice-enabled engineer booking solutions in its consumer energy business (British Gas) or within its smart home business (Hive).

Voice is such an obvious way to interact with “things” – “turn on the lights” or “boost my heating for 30 mins” works for me or my 84-year-old mum. Interfaces that are more natural mean less friction involved, less confusion, less training – they just work. The best-designed voice solutions also just take you there, don’t disrupt the flow.  Expect complex dialogue with some back and forth but simply turn your request or “intents” into actions.

What makes voice more compelling as an interface for the “smart home” is its ability to become a unifying, integration layer with technologies like Amazon Alexa and Google Home, allowing several different products to indirectly talk to each other through voice. A “turn off my lights” command won’t care if it’s a Philips Hue bulb, Belkin Wemo bulb or a Hive bulb – so long as they have all been set up to work with voice.

At the same time, we have also seen a rise in chatbots. All those tapping fingers, find it equally natural to tap out a command in WhatsApp, while watching TV or simply facepalming into their mobiles. Chatbots and pure voice are increasingly backed by impressive AI and natural language engines, sophisticated processors, and the power of the cloud.

It’s easy to overestimate where we are right now, as much of what is being designed is still using traditional commands on the back of voice request. They work really well for the smart home, picking a song, or setting the alarm, because the intent is clear and deterministic, and speed is of the essence. Any ambiguity to the request will only serve to frustrate the experience.

It’s clear, though, that voice is moving at a terrific pace and technologies like Google’s duplex, which has the capacity to convincingly mimic a human exchange like booking a restaurant reservation, is just the start of what will be a step closer to beating Turing’s cleverly devised “test of humanness”.

At Centrica we believe where voice could play a bigger future role is in the area of “peace of mind” applications, like “Assisted Living”. After all, voice is often easier to use for elderly people, who can use it to issue simple voice commands like calling for help, being prompted to take their pills and, no doubt, these will eventually develop into full conversational user interfaces (UIs) needed to one day play the role of a sophisticated companion robots.

But we are all aware it’s easy to get the experience wrong, particularly as user stories begin to demand more than simple voice commands to deeper conversational UIs. At Hive, User Centred Engineering and user experience (UX) is always our starting point. Understanding, through genuine customer research to establish the context, the current experience and what problems we are trying to solve. It’s then a case of applying solid UX principles, like reducing the cognitive load on the user, understanding what’s required of the conversation and when it’s necessary to be efficient and when it’s necessary to encourage dialogue.

Seb Chakraborty, global chief technology officer, Centrica Hive

Views from Wipro-Utility Week Technology and Innovation Council members

Nick Rutherford, 2020 IT investment programme director, Bristol Water

Amazon Echo (Alexa), Google Home and Apple’s HomePod, all voice technologies and yet another disruptor in the list of many technical disruptors that are emerging in today’s technological world. The market has grown since massively with an estimated over 40 million “voice-first” devices being sold by the end of 2017. Google’s says it sold “more than one” of its Google Home product every second between October and December 2017. With large growth forecast in 2018, some forecast that voice will dominate in years to come.

So will you be able to buy something using voice only and interrogate other services? Well you already can as Starbucks have been operating voice activated ordering for many months in some of their American shops which surely indicates the start of something bigger. Many others have already followed suit including online supermarket Ocado making it a first for UK supermarkets. You can ask Alexa to add garlic to your shopping order or check the status of your delivery.

So what about other sectors? Well adoption will be interesting; will people be comfortable talking to their little box as opposed to typing into a web page? For some yes and the adoption is likely to grow as more and more companies provide services supporting voice. However, Amazon etc and the companies offering voice enabled services need to ensure the customer experience is seamless and the technology doesn’t become a frustration. These products will need to evolve quickly as companies provide new services in order to gain consumer confidence and increase adoption.

Utilities and their respective digital customer strategies need to be considering voice as a new customer channel to serve. The trend is quite clear in terms of consumer demand, companies need to consider appropriate services to offer and be ready to support a future ‘new norm’.

Rob Saunders, interim challenge director – Prospering from the Energy Revolution, Innovate UK

When did you last catch yourself shouting at an inanimate object? Well, you might find yourself doing a great deal more of it in the future if voice control becomes ubiquitous.  Samsung, for example predict that talking to objects will become as commonplace as talking to people in the near future.  Is that really likely? And what might it mean for energy?

Talking is our natural way of communicating so it seems intuitive that voice control might minimise barriers to uptake of smart systems.  Right now, it’s just one way of controlling the actions of things around us.  While Alexa and her peers may be convenient to avoid having to press the ‘play’ button or open the door to see how cold it is outside, they hardly change the world for us, do they?  Well, not yet.  But what if voice recognition were the interface between you and a truly connected, intelligent home or workplace that learns what you want and how to deliver it?

The voice systems of the future will be able to understand nuance as well as content, so if you shout that you’re freezing, a different priority is understood than saying it’s a bit cold.  Whereas both would be interpreted similarly from turning the setpoint up on a smart thermostat, an intelligent system might tell you where it’s warmest in the house so you can thaw out quickly, and learn about the context of your mood and comfort in order to factor it into its future algorithms.

It’s this combination of data analytics, intelligence and nuance, that holds the key to a future where the voice-controlled, connected home seamlessly organises all aspects of your energy world.  In addition to automatically operating your energy assets to deliver warmth, light, power and mobility according to your need, a truly connected home will also optimise local generation, storage and demand flexibility to maximise your personal position in peer to peer, local and national markets to minimise your costs.

Now, maybe that’s a future worth shouting for?

Arnaud David, head of business intelligence – asset strategy, Affinity Water (former)

Utility companies are going through fast change, driven by the need to improve efficiency, increasing competition, rapid adoption of new technology, but more importantly, a shift in customer behaviours. In the past, utilities may have appeared to be struggling to match the leap forward in other sectors. However, they are catching up fast, and digital is playing a big part in this revolution, with the rise of the connected home. Companies like Google and Amazon are making their virtual assistants accessible to third party services and this is driving fast adoption and organic growth, supported by an explosion in connected devices.

Good customer service relies on sound and real-time information in the first place, which can sometimes be difficult to extract from operational systems and it will also take some time before smart networks become ubiquitous. At the moment, customers only spend a few minutes a year interacting with their utility suppliers to discuss their account or service.

There are three areas where voice-controlled interfaces can increase customer engagement:

Affinity Water has recently become the first UK water provider to offer customers a new Alexa skill for the Amazon Echo. When paired with their online “My Account” service, customers can ask for their account balance, find out when their last payment was and ask for water saving tips.  Customers with a meter will also be able to use it to submit a reading or find out their last one, helping them to stay in control of the water they use.

In future, it could be used to provide more tailored messages (no longer buried in a leaflet or website), by moving away from customer segmentation based on demographics, and focusing instead on behaviours and motivation, as well as property type. In time, it could even allow customers to control connected water devices and help reduce consumption, however we haven’t yet seen the level of adoption experienced with heating, lighting and security devices.

All of this requires an ability to manage and analyse big data to answer customer requests and questions in real-time, driving an even more proactive customer service from utility companies.