2024 Work Trend Index Annual Report: The Worrisome Microsoft and LinkedIn Report
All images via Dalle-2 and Microsoft Copilot. Some images may have been edited after their creation.
The 2024 Work Trend Index Annual Report from Microsoft and LinkedIn delivers some worrisome insights about the future of work. Most notably, it shines a bright light on the power of automation over common sense. For those of us who have worked in learning organizations, there is no more powerful outcome than an intelligent workforce engaged in actively improving processes, practices, and, most importantly, itself.
Self-taught individuals are bringing AI to work on devices and through browsers with little guidance, and most organizations have few plans for training. The overt focus on productivity in the Microsoft Work Trend Index does not provide the elevated perspective that generative AI deserves—the perspective required to recognize its transformative potential. The weak recommendations prove self-serving to technology interests rather than informative of broader opportunities for reinvention that could mitigate some of the failures inherent in e-mail and online meetings.
This post attempts to examine the report’s shortcomings, offer alternative perspectives on the data, and provide alternative recommendations aimed at engaging workers in co-creating their future.
Bring your own AI to work
I remember years ago sitting in a meeting while a young engineer was admonished for sending an email to a client via her personal email account. We were told that the attachment limits on the corporate email system made it impossible to send a large file via email. This was before the invention of filesharing services like OneDrive or Dropbox. Tools like FTP were cumbersome and locked down in an aerospace environment.
What wasn’t locked down were floppy disks. The young engineer copied the file to a floppy, took it home, and met her commitment to the customer by sending the company proprietary information from her personal email which had a larger attachment limit. She brought her own communication tool to work.
The Microsoft and LinkedIn 2024 Work Trend Index Annual Report states that 78% of people apply consumer or personally subscribed AI tools to work. That number says that people want to do work, and they want to do it well—and they see AI as a means to that end. The young engineer in the story above wasn’t trying to steal information or expose company secrets to the Dark Web; she was trying to meet a customer commitment for which she felt personally accountable.
People want to deliver value. They will find a way despite policy, even at the risk of their livelihoods. Why? Because the trust between customer and employee is often higher than the trust between employee and employer. Integrity often arrives from the bottom up. If AI can help do the right thing, it will be used even when it presents a different class of risk.
The broken psychology of AI
Interestingly, as much as AI has become The Skill that differentiates hires, 52% of those in the survey said they hide using AI for their most important work. Roughly the same percentage (53%) report that by using AI, they worry they will look replaceable. That’s a pretty damning indictment of management messaging. With 66% of the survey’s leaders saying they wouldn’t hire someone without AI skills, placing those people in an environment where they feel like those skills should become clandestine isn’t healthy.
Most corporations would do well to see AI as an opportunity not only for reinvention but for self-reflection.
Education receives short shrift
Education is imperative with the increasing pace of change driven by AI developments. LinkedIn has seen a 160% jump in AI training engagement among non-technical professionals. Only 39% of people in the survey received training at work, and only 25% of companies plan to offer training this year.
The lack of investment in skills isn’t new. Training budgets are often the first casualty of budget cuts when companies feel competitive pressure. While it’s easy to claim a budget win, ultimately, underinvestment in learning claims employee futures as its victims.
With such a high expectation for AI skills among new hires and a lack of talent that can meet competency thresholds, the best move for organizations is to upskill employees who already know their jobs. Yes, they may feel threatened by AI, but they also likely see learning AI as a personal advantage. Those worried about being replaced by AI will fare better if they know how to use it effectively more than their peers who don’t take up the challenge to master it.
In the short run, companies will not be able to hire enough people with AI skills. That leaves them two choices. The first is floundering, which is what many are doing now. By underinvesting in education, they force their employees to learn on their own, but that isn’t a sustainable solution. Employees who gain AI skills will not only be more valuable in their jobs, but they will also be more marketable to roles in other companies looking for talent with AI competencies.
And yes, there is always the fear that if the organization trains people they will take their training elsewhere. That thinking reinforces the emotional disconnect that most companies have with their employees. Broken reciprocal social contracts can’t be mended unless organizations start respecting and valuing employees. With labor shortages in an emerging skill area, the ability to not only empower employees with new skills but also create a working environment that encourages their retention will become a differentiator at a time of transition.
The second choice: Become a Learning Organization. As much as AI will rapidly imbue technical solutions with near-human intelligence, it will just as quickly expose human ignorance, perhaps even a lack of compassion.
The productivity problem
When I worked at Microsoft, I tried to break the cult of productivity. Since leaving, I have fought to create alternative views, most notably, my idea of the Serendipity Economy.
People are stressed. Microsoft rarely talks about its technology as being at the heart of the overload; however, too much e-mail and too many meetings run at the top of the list for most workers. Organizations don’t know how to get out of their own way. Too many unread (or under-read) e-mails lead to misunderstandings that often lead to meetings. Too many systems of record fail to reconcile, often leading to meetings. Too many operational systems that don’t reflect reality lead to meetings. Meetings lead to e-mails and conversation threads.
Turning information technology on a problem does not inherently increase productivity. It introduces new frictions and new tasks that take people away from meaningful work. This report says people only spend 40% of their time on creative apps; 68% struggle to keep up, and 46% feel burned out.
AI isn’t going to fix that, at least not if it is seen only as a productivity tool. But as a transformation tool, transformation will require people to step up to accountabilities and to create a new sense of what is important.
Zero inbox, for instance, is a meaningless metric. If much of my e-mail isn’t essential, my failure to delete it demonstrates proper prioritization. I am not spending time on things that do not warrant my time. Filing my emails may prove equally useless as powerful search engines now make it just as easy to find an e-mail in a cluttered inbox as in a well-maintained one. AI will make that even easier. What’s important isn’t having an empty inbox, but meeting accountabilities and delivering on commitments is important. I know of no performance review that has ever looked at the state of an employee’s inbox.
AI needs to be transformational for it to be meaningful. As long as Microsoft claims victory in minutes, the productivity problem will persist. This study classifies information workers as Skeptics, Novices, Explorers, and Power Users (the report’s terms) classified in part by the amount of time they save daily. The report offers some insight into Power Users feeling more in control and engaged at work, but it doesn’t set a benchmark for how these people felt before.
Power Users, from my experience, are motivated, self-starting learners regardless of the challenge. They love their work and engage in improving how they work as part of who they are. AI is just the latest tool for the current generation of motivated workers. At the beginning of the computing era, they were the ones who bought computers and brought them to work. They were the ones who mastered VisiCalc and knew all the dot commands for early word processors. Power Users were always going to jump on AI. There is no compelling insight about a class of worker who meets expectations.
Concerning, however, is the Harvard Business School report “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality” conducted with BCG where they found “beneficiaries of using AI are the bottom-half-skill subjects.” This makes sense. Those with less skills when given a tool that enhances those skills, will perform better.
Senior consultants didn’t improve as much because, well, they already knew what the AI was going to tell them, perhaps even disagreeing with it and having a firm foundation upon which to do so. Arguing with a productivity tool is never going to show well on a productivity study.
That study also fit right in with management consulting practice, which is to use lower-cost employees to do much of the work. And if those lower cost employees can meet the quality threshold with even fewer senior employees, so much the better. Here is how the report states that:
“Our results reveal significant effects, underscoring the prowess of AI even in tasks traditionally executed by highly skilled and well-compensated professionals. Not only did the use of AI lead to an increase in the number of subtasks completed by an average of 12.5%, but it also enhanced the quality of the responses by an average of more than 40%. These effects support the view that for tasks that are clearly within its frontier of capabilities, even those that historically demanded intensive human interaction, AI support provides huge performance benefits.”
What should organizations do
Unsurprisingly, Microsoft and LinkedIn offered little management guidance—a fault too often seen in technology companies as they analyze data. The answer is usually to use more technology. AI cuts down on e-mail reading, but It does not cut down on e-mail creation, which would be better. The company suggests that meetings gain value from AI. They do not recommend having fewer, more focused meetings with only the appropriate attendees or that accurate data reporting and trust increases the effectiveness of work more than meetings.
The document position, which I will restate in its entirety, makes a leap of logic too far:
“Overall, Copilot users edited 10% more documents in Word, Excel, and PowerPoint—the companies that saw the largest impact noticed a 20% increase. This may suggest that people are repurposing the time they save for high-value focus work like creating and consuming information.”
Shifting work is not proof of an increase in value. As I state in “The Serendipity Economy,” value is not derived from work associated with content until that content serves its purpose. Microsoft needs to teach its Copilots to discover the hidden threads of value derived from content, not the amount of time people spend creating or editing it. Content that creates no value has no value, even if it was created more efficiently.
The biggest problem with productivity is that any gains made likely result in either lower pay or eventual redundancy. Rarely do those who work in a job that sees huge productivity gains reap the rewards of those gains. Power Users survive because they can become Power Users in some other domain. For others, the impact of AI can be devastating. If it is to become as transformative as predicted (and we should take all speculative forecasts from a survey as being opinions from people who know no more about the future than we do), there will be significant displacements, which should force us to consider their fates inside of companies and as a society.
Invest in AI Competency: Of their recommendations, prioritizing training rings true, but not for the reasons stated in the report. Not everyone needs to be a Power User, but everyone needs to be aware of AI’s potential and have an appropriate level of competency to understand its impact on their jobs. Organizations need to become learning organizations not only because people need to continuously learn and enhance their skills, but also because the underlying knowledge is changing so frequently that someone hired as a competent master of ChatGPT 4 may find their skills don’t translate to other tools or newer versions.
Investing in AI competency also implies people need to have time carved from their work dedicated to learning. Perhaps those 30 minutes a day saved with a Copilot could be repurposed for learning.
Reimagine work: Much of this report presumes that work will be work, but AI will make it less burdensome by taking up some of the load. Organizations need to reimagine work to remove the obstacles to accomplishment with an eye toward AI’s capabilities. Using it to focus on minor productivity gains is anti-productive. Focus on eliminating activities and events, like meetings and emails, that contribute to the tediousness of work. At the same time, don’t just give lip service to giving people more time to contribute and create; spend time articulating pathways for that to happen.
Workers with strong AI competencies can participate in the co-creation of their work experiences. As I wrote in Management by Design, you can’t reengineer an experience that wasn’t engineered in the first place. We spend too much time ameliorating work practices that just happened rather than reimagining work through the lens of design. With generative AI, we have new tools that can move beyond incremental improvements to genuinely new ways of working.
Don’t overly constrain discovery with governance and policy: We are still in a learning phase. Companies need policies that will safeguard intellectual property, employees, customers, shareholders and leadership. But right now, that policy needs to be adaptive. Too many constraints will curtail experimentation, which may lead to reinvention. Most organizations are risk-averse, and their policies will reflect that perspective. Those who take more risks and monitor their investments effectively will likely realize the most benefit from AI investments.
Microsoft concludes that we have “arrived at a pivotal moment for AI at work.” Not yet. A moment, to be sure, but not a pivotal one. The generative AI tools remain disconnected from the process, without agency, within a mire of distrust and the emergence of AI shaming as people worry that, on the one hand, not knowing will prove detrimental to their careers and, on the other, that sharing how they use it in their core work will make them appear expendable. This is a chaotic moment of upheaval and learning.
As a scenario planner, I believe the outcome of AI’s impact remains uncertain. We may equally grow bored with it as to over-regulate it into irrelevance. We are just as likely to unleash it as a tool of autocracy as we are to nurture it as a tool for healing a broken world. All of those futures remain equally relevant. We are not at a pivotal moment but we are at a pivot. How wide the radius of the swing remains to be seen.
Those who engage, experiment and learn will find their way through. Those who try to preserve the status quo or wait for a competitor to show them the way through will likely be too late and too slow to reconfigure their niche for the next set of realities.
For more serious insights on AI, click here.
Thoughts or comments? Leave a message or ask a question in the comments section.
Leave a Reply