Mind the Gaps: Overcoming Friction in Implementing Workplace AI
When we look 5, 10, maybe 25 years into the future, we see workplaces where employees are using software that amplifies their efforts and improves their output. It’s a future where automation relieves staff of the menial tasks and drudgery that occupy so much of their day-to-day work experience today and allows them to focus instead on the fulfilling work they’re passionate about.
This vision is something we share with Matt Vasey, Myplanet Advisor and Director, Artificial Intelligence BD at Microsoft. He describes this future state as the “cognitive workplace”: an AI and data-powered way of operating that will fundamentally reshape certain key elements of how we work.
But to get to that cognitively-enhanced future, there will need to be some big shifts. Specifically, we need to build bridges that will take us from siloed, invisible, and difficult to access data to straightforward and simple data processing.
In a recent Forbes article, Hootsuite CEO Ryan Holmes talked about the potential game-changing future Microsoft was helping to shape when it announced a data-sharing partnership among Microsoft, SAP, and Adobe at the Ignite conference.
“The idea,” says Holmes, “is that data will, one day, be stored centrally and be able to flow smoothly between different systems run by each of these software giants.” As it stands now, software platforms tend not to have the ability to speak to one another.
“The different software systems used by brands generally speak different languages and store data in separate ways. The point-of-sale software in store may not integrate with the website. The marketing suite that serves you ads on social media probably doesn’t sync up with your purchase history. And the customer service platform may or may not know that any other system exists. And we haven’t even touched on the challenges of integrating real-life data from the Internet of Things or the sticky challenge of weaving together mobile and desktop info.”
By opening data channels, these software companies are helping set the foundation for the cognitive workplace that Vasey and others see on the near horizon.
“I think we’re set on the path to AGI — artificial general intelligence,” says Vasey. “We’re probably on a 10–20 year time frame for that,” he adds. He predicts that within the next two decades we’ll have intelligence capabilities powerful enough “to perform any intellectual task that a human being can”.
Joe Toscano, Founder of Design Good, echoes those sentiments. “When you hear Google focusing on Home or Amazon focusing on automating different services within your ordering system, these systems are just working in the background and you don’t have to go play with the screen, get glued clicking around. It’s just technology that works for you.” It’s what he calls “ambient computing” and it’s very much in line with Vasey’s thoughts on the cognitive workplace.
Because when it comes to workplace settings, the running-in-the-background, getting-stuff-done operations that come with a robust AI system alleviates a lot of the menial and time-consuming tasks employees are doing today.
We’re already seeing the impact of emerging technologies and IoT on a variety of workplaces, in particular in industrial settings. We see it in systems monitoring heating, cooling, and lighting settings, for example, which are being made hugely more efficient with the aid of “smart” technology.
We’re also seeing it on a personal level in those same settings. “You’re starting to see personal, Fitbit-like devices that measure fatigue and are able to provide workers with an environment that is safer,” notes Vasey.
These more industrial settings — where technology and humans have been in a close, symbiotic relationship and which have the infrastructure and resources to support further technological advancement, are a natural first point of entry for these AI-based updates. And inevitably, as these advances offer real improvements for workers’ safety, engagement, and productivity, they’ll begin to make inroads into offices, too. It’s already happening faster and with greater market depth than any tech advances before.
And that’s in part because these advances in tech are enabling their own growth and adoption. Historically, less tech-centric organizations were stuck with outdated and underperforming technology solutions because their infrastructure couldn’t adapt to meet the newest, latest, or greatest things on offer — at least not without enormous expense and uncertainty. But with AI, these companies are now able to make the switch sooner, because the new technology itself is flexing to meet the old systems.
“We see IoT and AI at and near the edge changing the way that employees interact with backend systems and making it easier for companies to adopt the technology over time,” says Vasey, adding that “there’s a whole set of partners that are going to be able to leverage the advances that AI and ML deep learning offer to automate business processes that happen in the cubicle or that information workers drive.” And that market, he says, is growing dramatically.
Microsoft’s 88 Acres campus has been leading the charge in corporate environments with their “Internet of Things meets Big Data” approach. They’ve turned their own spaces into a testing ground, showing how to make these changes at scale and how making these changes can have mammoth impact on efficiency and ultimately, bottom line costs.
Up to this point, smart technology and AI applications have had a focused, narrow-use view of things. They’ve been built to be highly specialized experts. But as the IoT expands and AI continues to evolve, gaining a closer resemblance to true AGI, we’re going to see less narrow use cases and much more flexible, adaptable intelligence software.
“We already have the super narrow AI applications today. But we’re beginning to see how we can broaden out that intelligence to take on more and more complex tasks now,” says Vasey.
Because the very nature of these systems is continuous growth, building on the data to harness the knowledge and create new knowledge from it, we’re going to see significant growth and adoption of new technologies in a way we never have before.
“There is a limit to our capability to train a machine learning model or neural network to deal with all the edge cases,” says Vasey, but those edges will continue to be pushed further and further out. The bigger limitation in the near term is what Vasey calls “hardware acceleration”.
“You need a fair amount of compute in order to learn some of these automation routines and there isn’t necessarily the hardware to do that today. But over time, that’s going to solve itself,” notes Vasey.
Hardware limitations are always a hiccup in the near-term, but long-term are easily solved. And the software limitations we once faced are quickly shrinking to be no more than mere hiccups as well.
There are other areas where the current gaps — those spaces between our current technological approaches in the workplace and the cognitive workplace of the future — are starting to shrink as well.
Take, for example, this knowledge gap we know is present today: there are more machines and automated assistants in use than people who can service them when they break down. We simply haven’t educated enough people yet, trained enough engineers and technicians, to be able to handle the widespread use of this kind of technology. But we will. And quickly.
Like the hardware limitations, these gaps will be overcome. In a few short decades, we’ll have graduated enough engineers, data scientists, and technical trades to overcome that problem and a host of others.
There are significant hurdles still to clear, but they no longer seem insurmountable or like some sci-fi far-off vision of a never-to-be-reached future state. We’re actively bringing that future state closer every day. The cognitive workplace will be here before you know it.
Interested in redefining what workplaces will be like when IoT becomes an integrated part of the action? Come work with us: we’ve got open roles in development, design, and product management.