Developing bespoke AI solutions is out of reach for many businesses but the cloud is providing an accessible path to adoption
When AI develops superhuman capabilities it often hits the news, like when AI beat a human fighter pilot in a simulated F-16 dogfight five times in a row.
However, the term ‘AI’ refers to a host of different technologies. Some of the more widely deployed subsets of AI in a professional context are machine learning, computer vision, and natural language processing.
Despite the potential for AI to impact society at large, many businesses are still at the beginning of their journey toward AI adoption, even those who are relatively advanced in their data journey.
“[One of the advantages of] the cloud is that you don’t have to bring in a lot of people to set up hardware for your AI and ML models”James Binford, VP, Cloud Security Solutions, US Bancorp
However, this may be starting to change. As more businesses harness the cloud and experiment with cloud-based services, AI and ML technologies are becoming more accessible than ever.
A recent Gartner study predicts that AI usage will increase by five times between 2019 and 2023, making it one of the most used cloud services. Furthermore, our own research from this year found that 63% of CDAOs have already started scaling their AI capabilities within their organizations, suggesting that this process is already underway.
For US Bancorp VP of Cloud Security Solutions James Binford, the greater accessibility of cloud-based AI makes it an attractive prospect.
“[One of the advantages of] the cloud is that you don’t have to bring in a lot of people to set up hardware for your AI and ML models,” Binford says. “I expect organizations like ours to take advantage of those offerings on a rapidly increasing basis.”
Taking the First Steps with Cloud-Enabled AI
AI technology could contribute up to USD $15.7 trillion to the global economy by 2030, according to a recent study by accounting firm PwC.
But despite the potential offered by AI, our research shows that 65% of enterprise businesses say that building a team with the right skills is a large or medium barrier to AI adoption.
That is just one element that could cause a business to think twice about building AI solutions from scratch, especially when there is no guarantee that such projects will successfully deliver ROI.
Data quality is another factor. Specifically, the ability to access enough clean data that is representative of the use-case in question. Although Electrolux is at a relatively early stage of AI adoption, Electrolux Director of Global Cloud and Hosting Services Kaveh Djavaherian is already putting cloud-based data architecture in place for future AI initiatives.
“I think that the use of [cloud] services to create data-driven or AI-driven analytics will accelerate,” says Djavaherian. “We are already creating data lakes and using cognitive analysis on them.”
Making sure that the data groundwork has been done to enable AI is also a priority for Siemens Healthineers Global Head of Cloud and Data Center Rohit Agrawal in the coming year.
“Once you have the data lake and the correct environment, then you can build on top of it,” Agrawal says. “The key is to have the right data at the right quality. So that is what we are focusing on right now.”
To maximize the chances of success for AI initiatives, cloud transformation leaders must make sure their infrastructure and data fundamentals are solid. However, as more companies take advantage of advanced cloud-based AI services, the pressure will be on to keep pace.
Cloud Security and the AI Arms Race
Many organizations, particularly those who hold large amounts of sensitive data like banks, have historically been wary about going too far too fast with AI.
Of course, there are many common roadblocks to enterprise AI adoption as we have discussed. There are also valid concerns about data protection and security, as well as the lively debate on AI ethics.
However, there could be a significant upside for financial organizations that can use AI effectively to tackle more substantive issues like credit risk evaluation.
Operating a machine learning model for credit risk evaluation is a thorny topic in the credit risk industry mainly due to the opaque nature of the AI processes involved. However, the benefits of adopting AI solutions for more advanced functions like credit risk evaluation may ultimately outweigh the risks.
“I think that the use of [cloud] services to create data-driven or AI-driven analytics will accelerate”Kaveh Djavaherian, Director of Global Cloud and Hosting Services, Electrolux
For example, the Chief Analytics Officer of banking group Banorte, Jose A Murillo, pioneered a data-driven credit risk evaluation program in his organization which helped increase the profitability of its credit card business by 25% in its first year.
“With data, you can be much more efficient,” Murillo says. “What kind of customers create a lot of risk? And how can you manage those customers better?”
James Binford VP of Cloud Security Solutions at US Bancorp thinks that competition amongst banks could be the spark that ignites the AI arms race in the financial services industry.
“I think the big driver of something like that is going to be someone else doing it effectively,” Binford says. “It takes that first domino to fall to set the others in motion.”
To allow cloud-based AI initiatives to flourish in enterprise businesses, it is essential that cloud transformation leaders do the necessary groundwork now. That means investing in the necessary data management, data governance and data security foundations to use the technology safely and efficiently.
Enterprises that aren’t doing so already are at risk of falling behind the pack. As our research has shown, the AI arms race is already underway. To the victor go the spoils.