If you’ve been paying attention to the latest surveys on artificial intelligence and, more specifically, to those focused on its adoption, you might have noticed a consistent answer: implementing AI is challenging and costly. In fact, many businesses trying to leverage this cutting-edge technology have always had this as one of their major objections to it, saying that it wasn’t worth the effort, given the little ROI they got out of it.
However, a new survey from Deloitte is suggesting that such sentiment is already changing. That’s because the survey’s results shed a positive light on the short- and mid-term future of AI implementation in business. How optimistic? Approximately 3 out of 4 executives respondents are confident that AI will be integrated into their enterprise applications within the next three years.
Machine learning, deep learning, natural language processing, and computer vision are all on the horizon. But what has changed for executives to make this strategic shift? Have they finally come around to the benefits of using AI, even if they aren’t readily available? Or it has to do with something larger?
Let’s explore the possibilities.
Developing Or Not Developing
The survey makes it clear that executives feel that AI will be more present in their business lives because of the possibility to buy pre-made AI-based products that can be easier to implement into their digital environments. Deloitte found out that half of those surveyed say they will either “buy all” or “buy more than build” AI solutions. Behind those numbers are all kinds of companies – from startups to large corporations eager to leverage the many forms of artificial intelligence.
So, as we estimated, Artificial Intelligence as a Service (AIaaS) model will be one of the trends to keep an eye in this crucial year. That’s because AIaaS can provide companies with the essential tools a company needs to assemble an AI-driven solution that’s more closely aligned with that particular company’s goals. Using this model, businesses of all kinds would be relieved of having to hire data scientists and AI developers, which are scarce and, therefore, very hard to find.
Problem solved, then? Not quite. The “buy rather than build” mentality still has its challenges. While it can provide a more immediate solution to the AI demands, it still needs a particular set of skills to understand and assemble. Given the offer based around AI, understanding the solutions that might provide a company with an advantage becomes essential. However, not everyone will be able to make an informed choice because of that requirement, so they’ll still need AI experts to choose the right AI solutions available in the market and for integrating it.
There’s another aspect to buying rather than developing. As it happens with all off-the-shelf solutions, companies jumping on this trend will have to make do with what AI providers offer them. Sure, there are huge companies behind AIaaS, such as IBM and Google, that are already offering specific high-end solutions to quench the companies’ thirst for AI. However, the offer is still somewhat limited and generic.
Those AI solutions available today can be the answer for what most companies are looking for – modernizing data infrastructure and automating tasks. However, as more and more businesses adopt the same approach, the need for more customized AI solutions will be more evident, especially for those companies wanting to get a competitive edge from their AI efforts.
So, developing isn’t totally out of the question – it’s postponed until entire industries standardize AI’s use.
AI Implementation: Easier But Still Challenging
There are two things worth remembering before jumping to conclusions about AI implementation shortly. First, this was just a survey (and one that was conducted in pre-pandemic times), so it doesn’t necessarily reflect a widespread sentiment. It’s better to take it as an indication of things that might be coming our way.
The second thing is closely related to the first: the fact that a group of businesspeople feels that implementing AI is easier today than it was in the past doesn’t mean that it’s a risk-free process. The same survey reveals that 56% of respondents say that their organizations are slowing AI adoption because of the emerging risks.
It’s worth highlighting that number, as it brings us back to reality: while AI might be easier to implement from a technical standpoint, it’s still filled to the brim with challenges. There are potential cybersecurity threats, the use of personal data in compliance with different regulations, the still raw nature of those regulations themselves – and that’s without mentioning the need to have AI experts conduct the implementation itself.
Under that light, we could say that AI implementation is more straightforward than it was the year before (and all the years before that as well). However, we’re still in a transition period for AI in the business world, and we still have a long way to go. Today, many companies are just starting to realize the benefits of using AI and aligning their processes and goals behind it.
Finally, let’s address the elephant in the room. I’ve mentioned it just in passing but it’s worth repeating: this survey was conducted in December 2019. In other words, the respondents’ feelings were coming from a vastly different context that didn’t account for dire economic conditions and a global impact brought about by a pandemic. As many companies are shifting off their strategies just to survive, the plans to implement AI will likely get postponed, even when AI could help those very same companies weather the storm and reinvent themselves into stronger versions.
All of these aspects converge into one virtually obvious conclusion: we can’t talk lightly about AI implementation today. We have to consider multiple factors at play during a year that’s widely different from others in recent memory. So, while AI might be technologically more accessible today, that doesn’t mean we’ll see an explosion in its adoption. If I were to guess, this marks the beginning of a trend that will blossom but will take a little longer to conquer the entire business world.
If you enjoyed this, be sure to check out our other AI articles.
