Forrester CEO: Lessons for Executives to Implement AI Successfully
The research firm kicked off its annual technology summit in Austin with advice for companies now in a race to adopt GenAI and other artificial intelligence tools.
George Colony’s enthusiasm for GenAI was clear at Forrester’s Technology & Innovation Summit North America -- and he believes the technology will tear down barriers between customers and companies.
Colony, Forrester’s chief executive officer and chairman of the board, said the company is doubling down on its call for clients to adopt GenAI and has added its own GenAI tool to the mix: Izola, which was demoed in the event’s floor show. The events speakers focused on three main GenAI-focused themes: AIOps (a newer term to describe the tools to monitor and observe AI performance), high-performance IT, and AI computing.
Colony, who last year created buzz when he predicted GenAI would bring about the end of the World Wide Web, said executives need to identify multiple AI implementation strategies. He used his own experience as an executive shepherding in new AI programs to Forrester over a very short period.
‘This is the easiest and hardest thing you’re ever going to do’
As Forrester developed its own large language model, the company had to go back to the drawing board several times. After being dazzled by what the product could do, the company realized there was room for improvement.
“We looked at it and realized, oh, this is only running at 65% accuracy -- not good enough for our clients,” Colony said. “So, the early part is euphoria, but then the hard work sets in. It took us about 12 months to move from 65% accuracy to 85% ... So, when you develop, it’s going to look easy at first, and then it’s going to get really, really, hard.”
Letting 1,000 flowers bloom
“You don’t know where the innovation is going to come from,” Colony said as he talked about his second lesson for executives. “You don’t know where the expertise is, and you don’t know where the ideas and where the coding capabilities are going to come from.”
Part of the problem is that emerging technology is evolving so rapidly, there’s a talent shortage of experts -- particularly in implementing GenAI. “You want to go and hire 100 of them,” he said. “They’re not out there. This is the next area of computer science. Everyone is learning how do to this.”
The answer to this problem is to get creative and multiply your efforts, Colony says. The early development team at Forrester was bolstered by the work of a college intern who was studying GenAI. “You want to look all around the company for where these ideas and the coding can come from. It could be someone from the marketing team, it could be a product technician … ”
‘It’s not that expensive … You just gotta do it.’
Colony said many clients are building GenAI projects on budgets of as little as $500,000. “We’re not talking about 10-million-dollar projects here, or $5 million projects. The models aren’t that expensive.”
Colony coined the “10 to 1” rule. He said companies should have 10 ongoing development efforts and have one in production by the beginning of 2025. “It’s fun to come to a Forrester event and hear about it, talk about it, read about it. But you just got to take that step to do this.”
He added, “It’s fun to go to the Taylor Swift movie, listen to Taylor Swift on Spotify, but it’s way more fun to go to a Taylor Swift concert. It’s the same with GenAI.
Projects need executive support
He said his own team needed his support during Forrester’s development process. While companies are facing budget constraints, IT leaders need to tap into executive decision-making to push AI projects forward.
“You can’t get this done without executive support … because this is scary,” Colony said. “You need executives to help the company understand what the opportunities can be from GenAI and not just the dangers. At a tech audience like this, I would say one of your jobs is to educate your executives and your CEO and your board in what this technology can do, and what the opportunities available will be. They don’t have to be experts, but they should be very functional amateurs with this technology.”
About the Author
You May Also Like
2024 InformationWeek US IT Salary Report
May 29, 20242022 State of ITOps and SecOps
Jun 21, 2022