Amazon is one of the most innovative companies in the tech sector right now, testing out ideas like artificial intelligence (AI)-powered self-serve grocery stores, autonomous delivery drones, and voice assistants in healthcare.
At its AI Innovation event on May 15 at the ecobee headquarters in Toronto, Amazon Web Services (AWS) spoke about how it is incorporating machine learning into its products and services, and why moving in this direction is so important.
“As we see the proliferation of data and more devices being connected to the internet, that creates an incredible opportunity for machine learning to be applied to solve business problems, to make different decisions, to go after really challenging circumstances, to address the needs of businesses, as well as consumers and citizens,” Eric Gales, director of AWS Canada, told CDN. “It’s an interesting area that Amazon has been investing in for two decades and it’s a big part of our business in terms of how it gets applied. Whether that’s using AI robots in warehouses or machine learning to develop better suggestions when you shop online, we’ve been focused on taking those capabilities and creating a portfolio of services to make it much more accessible to much wider range of applications.”
And while the Seattle-based tech giant is going all in on machine learning and AI, it recognizes that many of its partners need a boost. AWS has rolled out several machine learning API services to help them work more efficiently by providing them with an intelligent foundation to build their own products on. Amazon Lex, for example, can be used to build a smart chatbot that can engage with customers, while Amazon Rekognition Image and Video helps partners set up deep learning-based analysis of images and videos. So rather than partners needing to build and deploy the infrastructure needed to do image recognition, this API tool can be used by developers as a foundation.
“We’re focused on helping customers apply the technology in meaningful ways. We wanted today to showcase different examples of where machine learning can be applied so that customers can get some familiarity with its application, whether that’s voice recognition, or machine-computer vision across different scenarios. We’re seeing so many applications of machine learning across the world but a huge amount of interest here in Canada. There’s a lot of innovation being driven by Canadian customers using machine learning,” Gales explained.
Ecobee, for example, is a Toronto-based smart home device company that has embraced what AWS has to offer. It has started the process of moving all its data out of its current traditional data centre to the AWS cloud, and all future products will be cloud-native. The company also released a smart light switch – its first non-thermostat product – in March 2018 that is fully integrated with Alexa – a first on the market.
“We moved from being a single product company developing smart thermostats to one that can now work on parallel products that are AI-enabled, easily scalable, and can handle robust customer loads. None of this would be possible without a platform like AWS,” Jordan Christensen, vice president of technology at ecobee, said.
The British Columbia (BC) Institute of Technology is also taking advantage of Amazon Alexa. Bill Klug, the cloud computing option head and instructor at the university says that its school of nursing developed an Alexa skill that can read out a patient’s blood test results and diagnostic reports so that the nurses and students can stay hands free in the lab without risk of contamination.
Unbounce, a BC-based software company that builds and creates custom landing pages, is another company built on Amazon. The company uses AWS machine learning capabilities to tell its own customers whether their landing pages will be successful or not based on indicators such as the complexity of words on the page, how long the copy is, and how it can be tightened up.
Carl Schmidt, Unbounce’s co-founder and CTO, pointed out to CDN that websites with copy written at a sixth-grade level, for example, are more likely to be successful, especially in the real estate industry, and minimal copy drives success as well, with less than 500 words being the sweet spot for retaining audience attention.
“We’ve been on AWS since we first started in 2009 and having all of our data in the cloud on their servers has been great because it’s allowed us to keep costs down and scale up quickly. In our case, we use machine learning to provide real recommendations to our customers on how they can improve their websites. The biggest challenge we have these days is keeping up with Amazon’s pace and innovation. There are new tools coming out so quickly it’s hard for our developers to keep pace,” Schmidt laughed.
While some businesses are apprehensive about machine learning-enabled products and services, Schmidt said those feelings stem from a lack of information or misinformation.
“I think companies that are slow to adopt machine learning or AI just don’t even know what it is yet or don’t understand what it is. Everyone is hyping the technology up, saying it might steal our jobs or take over the world like the Terminator, but it won’t – it’s just about doing things smarter. Machines will become our counterparts, they won’t replace us,” he explained. “In our space – marketing – we have machines that can understand copy or recognize photos and videos, and then provide real recommendations that are effective. We used to have folks working on that because it was the only way; now machine learning does the heavy lifting and our people make sense of the data it collects.”
AWS Canada director Gales added that he’s seeing Canadian partners get on board quickly, countering the notion that Canada is often slow to adopt new technology.
“We’re seeing a lot of interest from the channel ecosystem and solution providers in helping to translate the AWS machine learning building blocks into real-life examples of applications that customers can use to help their businesses,” he concluded.