BOSTON — Adopting generative AI requires an upgrade in employee skills and internal trust, not just the quick adoption of new technologies, according to executive IT leaders.
IT leaders and panelists at IBM’s Think 2024 emphasized that the push to AI upskilling shouldn’t be limited to IT staff but should be pushed out more broadly to the entire enterprise.
They added that employees should feel empowered to experiment with the technology, such as at IBM’s Watsonx platform. Users should know there are well-defined safety guardrails protecting them, both in the software itself and through enterprise policies, the panelists added.
Having a future-looking set of policies will improve trust in the technology, according to Jikin Shah, global CTO and head of City National Bank technology at the Royal Bank of Canada.
We love every new, shiny thing, but we need to enable our partners and our end customers. Jikin Shah Global CTO and head of City National Bank technology, Royal Bank of Canada
Technology transformations rarely happen on time and on budget, he said. Providing employees assurance in a technology vision with clear examples of the benefits the new tools offer can ease that process, however, Shah added.
“We love every new, shiny thing, but we need to enable our partners and our end customers,” he said during a panel at the show. “In order for you to win in any technology race, you have to start with your people.”
Real world examples
Shah and the Royal Bank of Canada have taken a five-year view of how to implement new technologies, including GenAI, into the bank’s hybrid cloud, which utilizes on-premises, public cloud and private cloud infrastructure.
Those practices, experiments and knowledge base enable DevOps teams to know where it can push new code, he said, and has given the company the framework to start creating GenAI assistants for direct customer interactions.
Specifying what technology to use and when to use it is secondary to having employees who understand how and why it’s used, Shah said. He expects how and why questions to become more complicated as GenAI applications evolve, he said.
“We are very careful in how we approach AI,” Shah said. “We want to make sure the usage of GenAI is in a responsible manner.”
Carla Eid, senior vice president of data and AI architecture and governance at PepsiCo, said her data and IT teams were excited to jump on the GenAI bandwagon.
This eagerness started spiraling into complexity and other business challenges, she said, so she and IT leaders at PepsiCo eventually created a framework for vetting uses.
“We had teams from all over the world coming up with AI and GenAI ideas,” Eid said during a panel discussion. “We had to create a governance framework to assess validated and approved AI use cases against [our] responsible AI principles.”
PepsiCo uses GenAI and machine learning tools from both IBM and Microsoft, she said. Creating this framework made sure her managers understood what AI tools already existed at different parts of the technology stack and how those tools were used.
Providing that insight to all teams from a central location avoided duplicate efforts and unnecessary expenses, she said.
“Instead of every team having to govern AI models and incorporate responsible AI practices, we can do that centrally,” she said.
Executive planning with Watsonx
IBM’s Watsonx, the company’s hybrid cloud GenAI platform released last year, takes a more conservative approach to the technology compared with offerings from competitor cloud vendors such as AWS or Microsoft, said Tracy Woo, an analyst at Forrester Research.
Watsonx enables use of a hybrid or private cloud model for AI apps aimed at customers who want a higher level of data privacy, according to IBM. Services offered through Watsonx, such as its variant of the open source large language model Granite, also come with legal guarantees and protections, which enterprise legal teams may demand before adoption, she said.
IBM has a legacy in industries that lack experience in deploying cloud infrastructure, Woo said. Regulations in these industries, such as banking, might also curtail GenAI or other cloud technology adoption where Watsonx could better fit.
“[IBM’s customers] are not necessarily tech-savvy organizations,” Woo said. “They’re not working with born-in-the cloud companies.”
Unlike the push to the cloud in the past decade, enterprise adoption of GenAI has taken a slightly more sober approach in industries outside of the technology sector, said Krista Macomber, an analyst at Futurum Group.
The move to cloud helped many industries discover the criticality of data management and the responsibility of data security the hard way, she said, but GenAI is still viewed more as a curiosity among customers than as a necessary tool for digital transformation.
“People are aware of the risks and issues,” Macomber said. “We’re at least a step ahead compared to other technology transitions. When it comes to best practices, it’s still developing.”
Tim McCarthy is a news writer for TechTarget Editorial covering cloud and data storage.
