While artificial intelligence adoption is accelerating worldwide, the conversations at this two-day digital event highlighted that proficiency from company to company still varies significantly
Despite challenging circumstances around COVID-19, this year’s AI Champions, Online digital event highlighted how industry leaders are driving their AI initiatives forward.
“COVID-19 just accelerated [things],” observed Andrei Lopatenko, VP Engineering, Search and Conversational AI at real estate database company Zillow Group. “What would have happened in five years happened in five months.”
Market intelligence firm IDC reports that worldwide AI market revenues will exceed USD $156 billion in 2020, a 12.3% increase on 2019. Over the course of this two-day event, nine industry leaders shared how they’re maximizing the effectiveness of these AI investments and tackling AI adoption challenges in today’s business environment.
Shifting Customer Behavior is Changing AI Strategy
Customer behavior has shifted dramatically because of the global pandemic. In response, organizations have been quick to reevaluate their AI strategies and start innovating at a rapid pace.
“[What we learned is] that sometimes markets can suddenly change,” said Lopatenko. “For me, it was very important learning that engineering culture and machine learning culture should be organized in such a way that you are capable of adapting to these changes.”
Lopatenko argued that baking resilience into the cake helps an organization to be agile in the face of adversity. AI teams that can handle sudden change will outcompete those that can’t.
“You see what has happened in the market already,” he noted. “Some companies benefited from this change – they adapted very fast; they had a big increase in traffic – and some companies didn’t.”
Machine Learning Depends on Quality Data and Collaboration
Businesses are storing and generating more data than ever before. Ashish Bansal, Director, Recommendation Systems at streaming platform Twitch, argued that having this wealth of data is essential if you want to be sure you’ll have the right data for future AI use cases.
“It is worse to know that you threw something away when a use case arose than the other way around – that you have excess data but don’t have a use for it,” he said.
“The cost of storage is extremely small now and it’s falling rapidly,” he added. “So I think from a mindset perspective, you’d like to store as much as possible.”
Clearly, machine learning cannot be discussed meaningfully without addressing the data necessary to make it run. Yet, accessing and cleaning data remains a key challenge for many data scientists.
Chun Schiros, SVP, Data Science at Regions Bank, remarked: “Data quality is probably a universal issue, or a global issue, particularly to the financial [services] industry.”
With the right data in place, AI teams need to work cross functionally to enable AI and ML capabilities. Schiros explained that her team works closely with the bank’s IT, business and other groups to build AI products in a collaborative way.
She concluded: “The goal is to maximize the number of employees and people in the organization that are able to utilize the data to influence decisions day-to-day.”
Prepare for Five More Years of AI Innovation
New technologies such as blockchain and quantum computing are creating exciting possibilities in the AI space. Dr Satyam Priyadarshy, Technology Fellow and Chief Data Scientist at Halliburton, expects to see exponential growth for these emerging technologies over the next five years, spurred in part by the tumultuous nature of 2020.
“With multiple disruptive events, there has been a renewed interest by many organizations to actually transform,” he said. “[If we do this] we will actually be able to build disruptive engines for our businesses. Otherwise, many businesses will find themselves disrupted by other players.”
“It’s not just building the model (anybody can build the model), it’s about implementing it,” added Cleveland Clinic Director, Center for Clinical AI Aziz Nazha. “Then, [it’s about] showing the model actually driving benefit, whether that is improving patient outcomes or decreasing costs.”
Of course, universal AI is still a way off. There’s much work to be done to lay the foundations needed to operationalize AI in many organizations. But in the months and years ahead, we expect to see more and more AI leaders driving value across their organizations.
Brahma Tangella, Head of Analytics and Insights, US Patient Services at pharmaceutical company Takeda, concluded: “Data is the new oil. But maybe AI is the refinery and cloud is the pipeline.”