Houston, we may have a problem.
A human wrote this. 😊
Spent a bit of time this week with two primary research studies on the enterprise adoption and use of artificial intelligence. One from McKinsey, the other from Microsoft and LinkedIn.
Quick takes: McKinsey’s “The stage of AI in early 2024” confirmed that we’re in the earliest, testing-trial days of enterprise deployment. It also shed a bit of light as to where measurable value will be found. Microsoft's and LinkedIn's “2024 Work Trend Index Annual Report” revealed widespread individual and bring-your-own use of AI tools, with use often hidden from managerial sight – a two-part recipe for data protection and governance nightmares.
Taken together, they’re a wake-up call.
One which should rattle enterprise eardrums, especially in IT, HR, and throughout the executive corridor.
Let’s start with the McKinsey study, conducted online this year from late February through early March, with 1,363 participants across a global range of industries, company sizes, and functional specialties. Headline takeaways:
· 72 percent of respondents reported that their organization had adopted generative AI in at least one of eleven core business functions– a nice jump from the 55 percent reported the prior year.
· More noteworthy: only eight percent reported that the organization had adopted generative AI across five or more functions.
One business function. Per the Microsoft study (see below), might this be best understood as the reported, official number, akin to the measure of taxable revenues in an underground-economy?
According to McKinsey, the business function with the most AI traction was marketing and sales (checked by 34 percent of respondents), followed by product and services development (23 percent) and IT (17 percent.)
Two functions perceived by many to offer low-hanging gen AI fruit (software engineering and HR) lagged at 13 and 12 percent, respectively.
The most commonly reported use case? Content support for marketing strategy (copywriting by bot?), followed by personalized marketing (the same?)
Of interest: McKinsey asked in which functions (and by how much) generative AI use was leading to measurable cost reductions or revenue increases. The sample here had to be small, but the results were thought-provoking:
50 percent of respondents have experienced (or expect) significant cost reductions through the use of generative AI in human resources, with 15 percent overall experiencing or anticipating cost cuts of 20 percent or more.
53 percent of respondents have experienced (or expect) revenue increases through the use of generative AI in supply chain and inventory management. This figure must be understood as more of a forecast than a report, as only six percent of respondents noted current usage in that function.
Now let us go to the Microsoft-LinkedIn study, conducted in late February and early March 2024 by Edelman Data & Intelligence through an online survey of 1,000 or more full-time and self-employed knowledge workers in each of 31 markets. An additional sample of 2,800 U.S.-based knowledge workers was added to the mix.
Compare these results to the above McKinsey numbers, keeping in mind that Microsoft-LinkedIn focused on “knowledge workers.”
75 percent of Microsoft survey respondents reported use generative AI at work today, with more than half of those individuals starting their use of gen AI in the last six months.
The odds are that they are not all in marketing.
Here is where it gets interesting:
78 percent of AI users – by my math, that’s nearly three of five of all knowledge workers – are bringing their own AI tools to work. That’s BYOAI. By generation, the numbers are higher for Gen Z’s and Millennials, but even Boomers (73 percent) are in on it.
52 percent of AI users – by my math, that’s nearly two of five of all knowledge workers – are reluctant to admit that they use generative AI for their most important tasks. (Translation from survey-speak to cube-farm reality: no one is ever going to know.)
And,
53 percent of AI users – again, nearly two in five – fear that using generative AI on important tasks makes them look replaceable. (Again: no one is ever going to know.)
Flashing Red Lights Everywhere.
Let us go back to the McKinsey study for a moment. Officially, generative AI is being used in one or two business functions. Unofficially, it is probably being used all over the company, in all eleven functions.
Officially, it is under control, with no doubt an official, procurement-approved vendor list, a good list of governance guidelines, and regular reviews by senior management.
Unofficially, it might be running wild, loose, and out of sight, with God-knows-what entering hundreds, even thousands of corporate laptops, with – dare we think it – proprietary data squirting through the firewall.
Houston, might we have a governance problem?
We may also have an employment and talent drain problem.
Adding to the concern, again per the Microsoft study:
45 percent of U.S. executives are not currently investing in AI tools or products for their employees.
Only 25 percent of companies are planning to offer training on generative AI this year.
While 45 percent of knowledge workers fear that AI will eliminate their job, a similar number (46 percent) say they are considering quitting and finding new employment in the year ahead – a number higher than the 40 percent who said the same in 2021’s Great Reshuffle. (In the U.S., LinkedIn reports a 14 percent increase in job applications per role since Fall 2023, with 85% of listed professionals considering a new job this year.) In fact, 76 percent of Microsoft’s respondents said they need AI skills to remain competitive in the job market.
Hmmm. And,
55 percent of leaders are concerned that they’ll have enough talent to fill roles in the year ahead.
Wake-up call.
I am Jon Stine, 35+ years in retail business and technology. Most recently in conversational AI.
I read, I listen, I observe. I think, I write, I advise.
+1 503 449 4628.