- The Midas Report
- Posts
- AI Is Turning White Collar Work Cyclical and Investors Should Notice
AI Is Turning White Collar Work Cyclical and Investors Should Notice
3 min read.
Midas AI
August 11, 2025
As generative AI moves from demo to deployment, the most sheltered slice of the labor market is starting to wobble. That is not a headline from a hype cycle. It is a data point. JPMorgan senior U.S. economist Murat Tasci reports that for the first time, workers from non routine cognitive occupations now account for a greater share of the unemployed than workers from non routine manual jobs. He calls it an ominous sign and warns that the next recession could see AI tools induce large scale displacement in occupations that consist primarily of non routine cognitive tasks.
The shift matters because these jobs have been the ballast of modern recoveries. Scientists, engineers, designers, and lawyers barely dipped below pre recession peaks and often led prior employment rebounds. Meanwhile, since the late 1980s, routine jobs have been steadily automated away and still have not returned to pre Great Financial Crisis peaks.
Tasci’s point is not that AI will suddenly hit routine or non routine manual jobs harder. He argues AI does not pose much more additional risk to those categories, which require more physical personal interaction. The risk is concentrating where we least expected it. And because non routine cognitive workers now comprise nearly 45 percent of total employment, up from about 30 percent in the early 1980s, any sustained weakness in this cohort becomes a larger macroeconomic risk than in the past.
There is a counterpoint worth weighing. Investor David Sacks argues there is a clear division of labor between humans and AI. People provide context, write extensive prompts, and verify output. He says apocalyptic job loss predictions are as overhyped as artificial general intelligence itself, and that the old truism holds: you do not lose your job to AI, you lose it to someone who uses AI better than you. Both stories can be true at once.
Roles can remain while pathways shift. Evidence has been mounting that AI is limiting the number of entry level jobs typically filled by recent college graduates, which complicates how workers acquire the context and judgment Sacks says will matter most.
Where the Labor and Capital Flows Go Next
If you want to know where the labor and capital flows go next, follow the daily revenue work. In B2B sales, AI is no longer a sidecar. According to research, 82 percent of respondents already use AI to boost productivity and efficiency. Capgemini reports that 82 percent of surveyed organizations plan to integrate AI agents in the next few years. Gartner projects that by 2028, 60 percent of B2B seller work will be executed through conversational user interfaces powered by generative AI, up from less than 5 percent in 2023.
The feature lists are specific and operational. AI summarizes email chains and meetings, converts main ideas into tasks, automates pipeline approvals and follow ups, builds automation formulas from natural language, qualifies and scores leads, routes them by skills or territory, generates personalized scripts and emails based on client preferences and history, and powers dashboards with real time suggestions to advance stuck deals.
Platforms knit into the broader stack too, integrating with data enrichment providers like ZoomInfo, Clearbit, and Cognism, as well as demand generation and sales enablement tools such as Gong, Storylane, Outreach, Seismic, Highspot, and Showpad. Case studies claim material efficiency gains, like cutting manual work by 50 percent.
This is not gadgetry. It is a rewire of cost structures and growth levers. As seller work shifts to conversational interfaces and AI agents, unit economics change, sales velocity increases, and the human time that remains grows more valuable because it is applied to context and verification. Pricing is in scope as well. External research cited in the sales tech ecosystem points to AI led price setting expected to impact industries such as energy distribution and shipping.
Now connect that back to Tasci’s warning. If the most heavily AI exposed cognitive roles become more cyclical in downturns and less central to recovery dynamics, then the earnings streams tied to those functions will also behave differently. Businesses that lean into AI will not all move in lockstep, but the variance between adopters and laggards will widen.
The Investor and Operator Playbook
For investors and operators, the lesson is straightforward. Price in the speed and breadth of AI adoption where it is closest to revenue, and reassess labor assumptions in non routine cognitive categories. Evaluate whether a company’s sales stack already executes through AI for forecasting, prioritization, and outreach. Look for readiness to deploy AI agents, ability to integrate clean data, and a clear model for human oversight.
On the labor side, consider the exposure to entry level cognitive tasks and whether the organization has a plan to upskill workers into the higher context work AI cannot do. In portfolio construction terms, that is what an AI native posture looks like.
The Midas take is simple. Talent and capital are sprinting toward AI native strategies. Miss this shift and your portfolio may look analog. The better bet is to back teams building AI into core workflows and to train people to orchestrate it. That is where the next compounding advantage will live.