Last year everybody was talking about artificial intelligence and machine learning – technologies that are not just revolutionizing the delivery of customer experience, but also the way people are living their everyday lives. This year it’s expected that some of the most highly anticipated technologies in this space will finally become a reality.
To look ahead to four of the key AI trends we can expect to see emerge in 2018, I caught up with Brendan O’Driscoll, who was the founding CEO of a technology company called Soundwave, an Irish-based music discovery and personalization company. Soundwave was acquired by Spotify in 2016 and Brendan spent two years leading product for Spotify from the San Francisco office on all things AI, personalization, bots and everything relevant to this space.
During our conversation, Brendan outlined four key trends in AI and ML that he predicts will be big in 2018. Here’s a lightly edited version.
1. Voice interaction.
Brendan: Voice exploded at the CES conference this year. We saw all the major players adopting and integrating voice into their products and services. I think the major change that we’re going to see this year is mass user adoption driven by the likes of Amazon Alexa and the Google home devices. And from a software point of view, I think you’re going to see developers needing to incorporate voice interactions as much as touch interactions into their user experiences. So I think 2018 is definitely going to be the year of voice.
Patrick: So for people who just have a basic understanding of this they’ll know the likes of Alexa and Siri, where else might we expect to see this kind of technology pop up?
Brendan: Rather than just the generic platforms like Apple or Amazon you’ll see them pop up in your individual retailers. For example, target.com, nike.com, online grocery stores – everywhere where you purchase clothes, groceries, gadgets etc there’ll be a voice component. Through your favourite consumer experiences such as Netflix and Spotify, you’ll be able to say ‘hey play this song’ or ‘hey stick on that movie from the 80s with this guy as the lead actor’. So I think it’s just becoming more and more frictionless to search and discover using your voice rather than using any touch inputs or other ways of discovery.
Patrick: Do you think voice interaction is going to replace touch or is it just going to add to the channels that are available for people?
I think it’ll definitely complement not replace. There will always be a need for some elements of touch interactions like typing and using your fingers, but in many cases, it’s actually not the best way to interact with the user experience and I think voice will subtly start replacing some of the more frictional experiences.
2. Machine Vision.
Brendan: Machine Vision is definitely going to be another big one and I’m seeing some great companies leading the charge on this. For example, one company called LogoGrab have developed technology that can detect a logo in any image or moving set of images or video. Then we’re already seeing autonomous vehicles, they’re driven by machine vision which understands that a pedestrian is a pedestrian, a roadblock is a roadblock, and another car is another car.
And recently I’ve seen a really interesting play where a new start-up is focusing on the construction industry. It’s mapping out in real time what it sees on the actual construction site against what the plans for the building were, to make sure that they correlate 100 percent and there’s no mismatch between what was meant to be built and what is being built.
All of these are really good examples of companies that are leveraging massive amounts of data and ultimately creating vision within computers that can, for all intents and purposes, see and understand what they’re seeing.
Patrick: Where does this kind of advancement in technology leave human workers?
Brendan: Automation and artificial intelligence is not trying to replace the people, but rather reduce human error and push the people up the value chain so that they are more strategic in the equation rather than rather than lower down the value chain where a simple error could cost tens of millions of dollars on a construction site, for example.
3. Increased automation of menial tasks
Brendan: There’s a simple rule of thumb that says any task that can be done by human in a second or less can be and will be replaced by automation. So that could be something menial like confirmation of a calendar invite and other simple, basic tasks that can easily be replaced by bots. A lot of your day is spent on the kind of tasks that probably aren’t taking advantage of your skill set, and the aim is to wipe away as many of them as possible, giving you more time to focus on more strategic work
I’m a big fan of a scheduling bot that you can cc into your emails and the bot will coordinate with the person that you’re conversing with around availability. It knows your time, it knows their time, and it syncs everything up, doing all of the scheduling on your behalf. So rather than sending ten emails back and forth over four hours to organize a meeting, the bot can do that for you. Two great examples of these personal AI assistants are X.AI and Clara Labs.
4. Context-Aware Personalization
With the rise of Internet-connected devices we’ve got more and more sensors that understand more and more about us, and in 2018 they’re all going to connect together even more. So your fridge knows what’s in your fridge, and your thermometer knows what temperature it is, and your radio knows what day it is. This will all combine together and be able to give more inputs into the platforms and services that you use.
For example, your Spotify will know it’s raining and will offer you a raining playlist, and your training app will know that it’s sunny outside and will tell you to go do your sprints outdoors rather than indoors. All of these devices that are all in themselves smart can connect together and create a greater picture of the context that you’re currently in and create a better response to that with the services giving you what you want at that particular moment in time without you having to ask.
Incorporating AI into your own organisation
Patrick: For companies reading this because they want to start automating some of the systems and processes within their own organisations, how would you suggest they approach it?
Brendan: Well I think companies should approach it as a classic ‘Partner, Build, or Buy’ approach – one of those three options. First off, to truly understand the power of some of the technology that’s available, I would suggest that they look at partnering with technology firms that can offer services on a kind of SaaS-based model. Once they have a greater understanding of the advantages I would suggest bringing that in-house and either building your own resources that can do this type of thing or else buying a company that can expedite your ability to develop your own automation resources.
Patrick: Brendan, thanks so much for your time and insights into this fascinating space.
If you like this kind of thing, we reckon you might also like: