Alan Hartwell, chief technology officer, Capita Software
We are seeing ever more people use virtual assistants like Alexa, Siri and Google Assistant, and trust the results. Voice assistant technology is now part of normal life, and people will increasingly expect it in their business life too.
In 2019, automation – from Robotic Process Automation to Autonomous Vehicles to Automated Intelligence – will finally become truly mainstream. As the resulting data sets from business usage become larger, automated analytics will exponentially improve the results and deliver compelling business ROI.
1) The capabilities of virtual assistants will grow rapidly, thanks to a surge in available data
As more and more people use voice assistants, the data that’s collected will grow and the outcomes become better. It’s a virtuous circle. The bigger the data sets, the better the performance; the better the performance, the more people use voice assistants; the more people use voice assistants, the larger the data sets become, etc, etc. It takes large data sets to make those significant improvements happen, but by 2019 we’ll be starting to see them.
2) Analysis of that data will make them smarter
The reason the results improve is that you can analyse the voice recordings to match successful outcomes with unsuccessful ones.
You can track back over the conversation and learn from it, and that means you can make better predictions over where a future conversation is going to go.
For example, we’re already starting to see in financial services that it’s possible to listen to a call from a customer and tell, from the words being sampled, that the right advice isn’t being given. It’s pretty niche at the moment, and you need large data sets to validate and refine it, but I can see systems like that becoming more widely adopted.
3) They’ll begin to handle more complex enquiries
Because voice systems are going to become much better at interpreting what’s being asked, I think we’ll also start using them for more complex enquiries. Take “am I on the right gas tariff?” for example. If we were trying to answer that question ourselves, we’d first check our account, then go online to look for better prices, then maybe use a site such as uSwitch. But when one phrase can activate all of those processes for us, that’s when it really becomes worthwhile. When something that’s easy for a human can also be done by a machine, it’s an amusing novelty. When something that is complex for us can be done easily by a machine, that’s real value.
4) Businesses will find new ways of using them to enhance productivity
Increasingly we’re going to be hiring people who don’t want to go on a three-day training course to use all our internal systems. They already live a life that works pretty much out of the box and they expect their working life to be the same. So there are huge productivity savings waiting for companies that can use virtual assistants to do away with the need for that training. Just press this button on your PC and ask a question – how do I speak to HR, how do I change my address on my personal records, how do I do my expenses…? That’s a business use I certainly see expanding.
5) But the issue of how much data they gather on us is not going away
As corporations gather more and more data about people, you begin to get into the area of predictive analytics on a personal level – personal management based on observations of your behaviour. Not everybody feels comfortable with that, certainly not in the Cambridge Analytica/Facebook sense, and it’s a subject I suspect will generate even more intense debate in 2019. For some, it feels weird to have an app that says: “we notice you’ve been making lots of payments in fast food restaurants this week. Are you sure that’s good for your health?” But it does point the way to the future. Useful, proactive prompts that are not being sold to massive corporations to make money out of you, but delivered just to you, for your own benefit. Yes, I’d look forward to that.