BOSTON — Enthusiasm may remain high for organizations around the prospects of artificial intelligence (AI). However, people still have concerns around data integrity and integrations. According to an Accenture survey on European government organizations, more than two-thirds (71%) foresee trouble procuring data integrity and processing capabilities. Even more (84%) cited challenges in adapting AI logic and reasoning to their industry context; another 81% said they experienced challenges integrating AI technologies into their back-office operations.

CMSWire was on hand at last week's AI World in Boston at the Seaport World Trade Center to catch up with those wanting to learning more about adopting AI into the enterprise. 

Some of the prevailing themes at the conference included the recognition that AI in the enterprise is still a nascent venture, and that AI won't produce tangible outcomes without a good foundation of usable data. Here are some other tales from attendees at the conference.

Making Better Use of Data

Traci Bartos, director of product supply IT operations at Bayer, said her challenge is not the collection of data from the "many different forms and systems" but rather making good use of it. "We need to connect that data to take action," Bartos told CMSWire. 

headshot of Traci Bartos
Traci Bartos

"We're here to learn about what AI technology offers in that regard. Our current systems are data collectors. A lot of what we do today we do in an ERP system. How do we make use of what we do with our current system? How do we make better use of data?"

AI Investment is a Skill Set Shift

Lisa Frank, Bartos' colleague and director of research and development for IT operations at Bayer, said an investment into AI is about finding the right skill sets among people, the right technology vendors and useful partnerships with third parties and within her own company. "A lot of what we're doing is to see and learn if we can make connections and help accelerate our business," Frank said.

Headshot of Lisa Frank
Lisa Frank

Frank calls the effort a "skill set shift." What differentiators are "going to make us have a competitive advantage and to be able to keep up?" Part of that effort is connecting divisions within her own company because some departments are more advanced than others in certain areas with their data platforms. "And we're learning and trying to understand what they're doing because we want to get those synergies with the other divisions," Frank added. "It's all about the people at the end." 

Related Article: Why Business Intelligence Has a Role in the World of AI

Academia's Recognition of Data

Speaking of academia, some universities are recognizing the growing interest in AI and data science. Jackie Klatt, education recruitment and admissions specialist at the University of New Hampshire, was on hand at AI World to tout online graduate programs like Data Science. The coursework includes SQL and NoSQL, Python and Git and predictive modeling. Why the rise in such courses?

Headshot of Jackie Klatt
Jackie Klatt

"Everybody regardless of what they do in their job is working with data in some way," Klatt said. "We have people who only work with data occasionally and they just want to understand this a little bit more. They see all those numbers and all those numbers and all those Tableau dashboards and they want to know what to do with it and how do they get there. They want to be really good at data visualization. We also have people who want to become data scientists."

Learning Opportunities

Do You Have the Data to Implement AI?

Who's winning in AI? Organizations like Apple and Google are finding ways to win with AI because they "have the data to support it," according to Mike Salzarulo, systems engineer at Centauri. "That's the biggest problems people are having," he said, referring to leveraging good data. "They're still working on handwritten forms. That's difficult because computers don't understand handwriting."

Mike Salzarulo headshot
Mike Salzarulo

Many organizations don't recognize they don't have the data to work with AI or machine learning. Another problem, according to Salzarulo? Executives with "wild imaginations" that AI can be infused across an enterprise and leveraged immediately. "We're not anywhere near that," he said.

Related Article: Artificial Intelligence Threats and Promises

Creating an AI Center of Excellence

Many practitioners are focusing on organizing AI as a practice in their organizations, according to Gigi Freeman, client partner at Catalytic. The problem she hears? Many of those practitioners, charged with AI implementations by their boards of directors, "immediately stand up a Center of Excellence for AI but then it's very siloed and not necessarily achieving what they wanted to achieve."

Headshot of Gigi Freeman
Gigi Freeman

Should the AI Center of Excellence be siloed from other departments? Do you have a Center of Excellence that provides governance to allow other people in the business to experiment with AI in their various domains? "That's been an interesting discussion to be a part of," Freeman said. "And I think at the end of the day, you need to empower your employees to let them experiment, but you do need to have that governance and infrastructure."

AI Preparation, Explainability Musts

One tech company offering AI software said AI implementations need to recognize humans still play a major role. Ariel Elizarov, CEO and founder of Lazarus, which helps doctors detect cancer risks in patients, said the first step in an AI implementation is being prepared to work hand in hand with these AI systems. His company's technology acts like clinical decision support for doctors so they "can make better clinical decisions."

Ariel Elizarov headshot
Ariel Elizarov

Another critical component of AI in an enterprise, especially in healthcare? Explainability. Elizarov finds this incredibly important and ensures explainability is built within company's deep neural network systems. "You should not be working with the black box if you're in the healthcare space," he said. 'Because if something bad happens, the doctor needs to be the person to (address it) at the end of the day. You don't want to trust it (without the doctor), at least not now, because it's still nascent."