In this environment, COOs of Fortune 500 companies overwhelmingly tell us that they have a neutral outlook over the next one to two years, focusing on strategic rightsizing decisions on labor and capital investments. The EY Center for Executive Leadership hosted these COOs for a virtual roundtable, touching on today’s economy along with GenAI use cases and trends in how executives are thinking about the technology.
Here’s what COOs should know and key takeaways from the discussion.
Expect rebalancing in the economy, setting up a year of growth in 2025
In the US, economic growth will continue along 2023’s modest pace, driven by somewhat slower labor market momentum, said Lydia Boussour, EY-Parthenon Senior Economist. Although some activity is cooling, she sees no signs of a global recession — rather, the economy will continue to rebalance, setting up a year of stronger growth in 2025.
Dynamics in services industries and the labor market have been keeping inflation a bit elevated in the US more than in other geographies, positioning the Fed to cut interest rates only by 50 basis points this year, starting in the summer, Boussour predicted. Job openings are close to pre-pandemic levels, and labor supply has climbed as well: many people who left the labor force are now re-entering, coupled with increased immigration.
The upcoming presidential election weighed heavily on the minds of COOs. Boussour said that her group modeled scenarios for a potential change in control of the White House, primarily in two areas:
- Trade policy
Additional tariffs floated by former President Trump, not just on China, would increase the cost of traded goods and act as a significant drag on GDP growth. For his part, President Biden has also recently introduced tariffs on some Chinese goods.
- The Tax Cuts and Jobs Act, which is set to expire in 2025
“Congress will be focused on this no matter who wins,” Boussour said. If nothing is done and the tax cuts expire, as is more likely Democratic administration, the impact on the economy would also be sizable — a “mini fiscal cliff.”
Our COOs are approaching a year of rebalancing with cautious optimism. “We’re being more pragmatic on enterprise initiatives in digital transformation and investment,” one COO in the chemicals sector said. “We’re making sure we’ve fully implemented and realized value before getting into something new. The pace and number of projects are less than before. On the flip side, we’re taking a three-year view on investment roadmaps instead of an annual view, and making decisions on a quarterly basis rather than in annual budgeting plans.”
Another agreed that with global instability and the upcoming election, the goal is to hold costs as low as possible and put off larger enterprise programs over 18 to 24 months. “We see 2025 as the first time period to connect to 2019 — the end of the pandemic and its aftereffects,” this COO said. “We’re trying not to do too much as we see where interest rates come down.” He added: “Labor pressures have reduced, and we’re doing well on shopfloor hiring, but we’re doing more work on campus now than ever for engineers.”
Build out your AI capabilities as “agenic AI” takes shape
The conversation on labor and productivity provided a natural transition to artificial intelligence and developments in GenAI. Today’s GenAI use cases are often divided into two groups: reactive chatbots that help workers reduce their everyday toil when prompted (much like today’s GPTs) or proactive copilots that work collaboratively (such as by automatically providing a transcript of a call, for instance), with the human worker ultimately making decisions. But Dan Diasio, EY Global AI Consulting Leader, introduced a third option: “agenic AI,” or “AI agents.” These AI agents can independently run multiple steps of a process to deliver outcomes.
AI agents run functions for multiple steps of a process independently, generating capabilities to complete a task on their own, and the next task, until the ultimate outcome is delivered, Diasio said. If you prompt a GPT to write an essay, it spits it all out automatically — not how a human would do it, or at least not for an optimal result. Perhaps you would outline your thoughts first, do research and gather feedback. An agent makes this approach more feasible in the realm of GenAI by initiating a thoughtful process toward a better output and not just completing a task.
Initial GPTs were accurate roughly 65% of time, Diasio said. But through agenic techniques, the outputs become roughly 95% accurate, although, at this point, it’s more costly as well. “It starts to address hallucinations and false retrievals,” he noted. “Those problems are why finance and supply chain have been slower to take this on. And studies suggest that performance decreases in a copilot situation, although that’s not fully settled. Companies are creating clear responsibilities between AI and people, and they’re not conflating the two.”
One of the participating COOs indicated the strategic embrace of agenic AI as a pivotal direction of where they were headed as an organization yet stressed the importance of solidifying core infrastructures such as data and ERP platforms. Their investment strategy was to enhance capabilities and drive value, with a rigorous “shark tank” approach to digital innovation led by a dedicated team.
What is your view on the potential of AI, including GenAI, to impact supply chain and operations?