Understand AI
getty
Is artificial intelligence living up to all the promises made by its proponents? Business leaders remain interested in generative AI, but their enthusiasm is waning, according to a recent survey conducted by Deloitte. says. 67% and 57% of senior executives and directors have “high” and “very high” interest in generative AI, down 8 and 6 points, respectively, from the first quarter of this year.
“This is likely because many genAI initiatives are still in pilot or proof-of-concept stages,” said Jim Rowan, head of AI at Deloitte. In total, the majority of respondents said their organizations have less than 30% of their generative AI solutions in production today. “This makes it difficult to see the market impact of these initiatives,” Rowan said. “Until these initiatives mature and scale, it is difficult for organizations to know whether they are meeting initial expectations.”
In short, we are still in the learning phase. “In the industries I serve, everyone is engaged in large-scale, global experiments to see how generative AI and related technologies can create lasting business value. Michael Umlauf, TransUnion’s senior vice president of data science and analytics, told me. “As experiments progress and we continue to learn more, including how to properly manage these tools and related systems, expectations will naturally begin to fall and become more firmly grounded in reality.”
Companies are also continuing to learn about the common problems that arise with AI. “Companies are facing all kinds of constraints, from new governance requirements to a lack of skilled talent,” Umlauf said.
At the same time, Umlauf added, “We are witnessing a real willingness from our business partners and stakeholders to engage and learn from each other’s successes and failures.”
AI may be overrated, “but for those who understand how it applies to their business, it can be transformative,” says software developer said Courtney Machi, vice president of Andela, a global job placement network for people. “Companies that put resources into understanding how to leverage resources into their business are moving forward.”
Part of the challenge, Machi says, is that “many people struggle with how to get started and use cases, and don’t have the right people on their teams to work on solving problems with AI.” Some people stick to internal productivity use cases with co-pilots like enterprise technology, which can be beneficial but difficult in terms of tying into ROI. ”
Another factor yet to be determined is measuring the return on investment for AI. “It depends on how realistic their expectations were from the beginning and whether they took the time to define clear goals, performance indicators and success criteria upfront,” Umlauf said. . “Many of the early successes have come in the form of increased productivity for knowledge workers, and benchmarking against existing practices should be a fairly straightforward task.”
Current generative AI tools are versatile because they can “help programmers write better code faster, create early drafts for content authors, and quickly extract insights from long sets of documents.” It really shows sexuality,” Umlauf added.
ROI is for companies that are able to reduce costs by automating tasks traditionally performed by humans, or grow sales with differentiated products, and that 20% of AI output is accurate ” said Mr. Machi. “It goes back to whether companies are willing to invest the time upfront and whether they’re willing to take some risk.”
When implementing AI, most organizations “start with tactical benefits, such as improving existing processes or reducing costs,” says Rowan. “Essentially, they have achieved low-hanging fruit to unlock immediate value while building their knowledge, experience, and confidence in AI. But now they are looking to scale that value many times over. We’re trying to scale up, but that comes with other challenges.”
Beyond simple productivity, more work is needed to measure benefits in more ambitious use cases and “demonstrate improvements compared to existing solutions,” Umlauf said. Ta. Established solutions “have often been honed over many years to high levels of performance and accuracy on structured datasets.”
As more AI projects move “beyond proof of concept and successful projects are rolled out across sectors, we will see further impact and return on investment,” Rowan said. “Currently, executives may be making decisions based on fear of missing out, meaning that maintaining executive interest and buy-in as hype levels off requires measurement. It becomes an important element.”
Realizing the benefits of AI “requires selecting the most promising use cases, diligently measuring AI performance and alternatives, and engaging in a process of continuous improvement,” Umlauf said. Masu. “In a narrower sense, companies can consider their AI projects successful if they can clearly and consistently demonstrate superior performance compared to benchmarks and achieve widespread adoption. One sign of success is that AI will recede into the background and become another tool that helps humans achieve their goals more effectively and efficiently, so we can look forward to the next wave. I guess.”
