The history of technological disruption to labor markets has a recurring structure: a new general-purpose technology arrives, analysts predict mass unemployment, the near-term impact is more concentrated and complex than predicted, and the long-term impact is more structural and far-reaching than the optimists hoped. Steam power, electrification, computing, and the internet all followed versions of this pattern. Artificial intelligence is following it too — but with a critical difference. Previous technological waves primarily disrupted physical or routine cognitive labor. The current AI wave is demonstrating meaningful capability in exactly the domain that previous automation left untouched: professional knowledge work.
This distinction matters enormously for investors, for the companies building in this space, and for the professionals whose labor markets are being restructured. At BeMoreeDriven Capital, we have spent the past eighteen months developing a detailed view of how AI is transforming professional labor markets — not at the level of headline narratives about robots taking jobs, but at the level of specific workflow disruption, business model restructuring, and the emergence of new infrastructure layers that will define the economics of expertise for the next decade.
The Data Infrastructure Prerequisite: Scale AI's Role
Any serious analysis of AI's impact on labor markets has to begin with the infrastructure that makes AI capabilities possible. Scale AI, which raised a $1 billion Series F at a valuation of approximately $7.3 billion, occupies a position in the AI infrastructure stack that is often underappreciated by analysts focused on the visible surface layer of generative AI applications. Scale is fundamentally a data labeling and AI training infrastructure company — the picks-and-shovels business that every frontier AI model depends on for the high-quality training data that produces useful capabilities.
Scale AI's business illuminates something important about how AI labor market disruption actually works. The company employs or coordinates a global network of data labelers — human workers who annotate images, transcribe audio, evaluate model outputs, and perform the highly specific data preparation tasks that enable AI training. In other words, Scale AI's business model involves paying human workers to perform the tasks that, over time, the models being trained on their output will increasingly be able to perform autonomously. This is the paradox at the center of the AI labor market story: the development of AI requires significant human labor investment now, in exchange for a reduction in human labor requirements in the future.
The implication for investors is significant. The transition from human-intensive data preparation to AI-assisted and eventually AI-automated data preparation will restructure Scale AI's cost structure in ways that could either dramatically increase margins or require fundamental business model reinvention. The company appears to be actively navigating this transition — investing heavily in automated data quality evaluation and model-assisted labeling that reduces the human labor required per unit of training data produced. Whether Scale emerges as a durable infrastructure layer or faces disruption from the very capabilities it helped create is one of the most interesting structural questions in the AI ecosystem.
For the labor market, the Scale AI story represents a pattern that will repeat across many sectors: AI development creates new categories of work (data labeling, model evaluation, AI quality assurance) while simultaneously developing the capabilities that will eventually displace those workers. The transition period — where both the new work and the old work coexist — creates investment opportunities in the infrastructure that manages that transition. Understanding these transition dynamics is central to how we evaluate AI-adjacent investment opportunities.
Professional Knowledge Work: Harvey and the Legal Sector
If Scale AI represents the infrastructure layer of the AI transformation, Harvey represents its application layer — specifically, the disruption of one of the most sophisticated categories of professional knowledge work: legal services. Harvey, which raised a $100 million Series B at a valuation of approximately $1.5 billion, is building an AI platform specifically for legal professionals. The company's technology draws on large language models fine-tuned on legal texts and workflows to assist lawyers with contract analysis, due diligence, legal research, regulatory compliance, and document drafting.
The legal sector is an instructive case study because it represents a profession that has historically been remarkably resistant to technological disruption. The complexity of legal reasoning, the judgment-intensive nature of legal advice, and the high stakes associated with legal errors have made law one of the most durable bastions of high-cost human expertise. Law firms routinely bill $500–$1,500 per hour for senior associate and partner time, and the leverage model of the traditional law firm — where junior associates perform high-volume, relatively routine work at rates that fund partner compensation — has remained essentially stable for decades.
Harvey is attacking exactly this leverage model. The company's platform can perform contract review and due diligence tasks that would typically require hundreds of junior associate hours in a fraction of the time, with an accuracy rate that compares favorably with human performance on well-defined legal tasks. This does not eliminate the need for senior legal judgment — the interpretation of ambiguous contractual provisions, the strategic advice about negotiating positions, the exercise of judgment about risk tolerance — but it dramatically reduces the volume of routine, high-cost human labor required to produce that judgment.
The business model implications are profound. Law firms that adopt Harvey's platform can either reduce their headcount (and costs) while maintaining output, or maintain their headcount while dramatically increasing output per attorney. The former path is deflationary for legal employment. The latter path, at least in the near term, increases the productivity and potentially the compensation of the attorneys who remain — because they are doing more leveraged, higher-value work while the AI handles the routine volume.
What Harvey's trajectory suggests for the broader labor market is a principle we are applying across our sector analysis: AI does not uniformly eliminate jobs. It restructures the leverage model of professional work, concentrating value in high-judgment, high-relationship roles while reducing the value of high-volume, routine professional tasks. The winners in this restructuring are professionals who can effectively leverage AI tools to multiply their output. The losers are those whose value was primarily in performing the tasks that AI can now perform at lower cost.
The Expertise Democratization Effect
There is a dimension of AI's labor market impact that receives less attention than the displacement narrative but may ultimately be more economically significant: the democratization of expertise. When AI can perform meaningful portions of the work previously done by expensive human experts, access to that expertise is no longer constrained by the ability to pay expert rates. This has potentially enormous implications for the economic geography of expertise and for the businesses that can now access capabilities previously available only to large, well-resourced enterprises.
Consider the implications for small and medium-sized businesses. A company with 50 employees that previously had access to legal counsel only for major transactions now has access to AI-assisted contract review for routine agreements. A startup that previously could not afford sophisticated financial modeling can use AI tools to produce analysis that approaches what a boutique advisory firm would have charged six figures to produce. A healthcare provider in a rural community that previously lacked access to specialist diagnostic expertise can now leverage AI-assisted diagnostic tools trained on data sets that no single specialist could review in a lifetime.
This democratization creates new markets for the platforms that deliver democratized expertise, but it also creates competitive pressure on the traditional providers of that expertise. The incumbents in every expert services sector — law, accounting, consulting, medicine — face a structural challenge: the aspects of their value proposition that were previously defensible because they required expensive human scale are being commoditized by AI. What remains defensible is the judgment, relationship, and accountability components of professional expertise that AI cannot yet replicate.
For investors, the key question is not which expert services businesses will be disrupted — the answer, broadly, is all of them — but which companies are building the platforms that deliver democratized expertise and capture the resulting economic value. Harvey is one example. The legal AI space also includes Clio, Contract Counsel, and a range of more narrowly focused tools addressing specific legal workflow components. Each represents an attempt to build the infrastructure layer of democratized legal expertise.
The Reskilling Problem and the Investment Opportunity It Creates
Every major technological transition creates a reskilling problem: the workers whose existing skills have been displaced need to develop new capabilities to remain economically productive, and the institutions responsible for providing those capabilities — universities, community colleges, corporate training programs — typically adapt more slowly than the labor market demands. The AI transition is generating a reskilling challenge that is broader, faster, and more complex than any previous technological transition.
The scale of the challenge is reflected in the labor market data. A 2024 McKinsey Global Institute analysis estimated that between 2024 and 2030, approximately 12 million workers in the United States alone would need to transition to different occupational categories as a result of AI-driven automation. The required transitions are not simply from lower-skill to higher-skill work — they are often lateral, from one type of cognitive work to another type of cognitive work, across occupational categories that do not have obvious training pathways.
This reskilling challenge creates a substantial investment opportunity in workforce development platforms. The companies building effective reskilling infrastructure — tools that can identify skill gaps, develop personalized learning pathways, and deliver training efficiently at scale — are addressing one of the most significant labor market problems of the decade. The companies we find most interesting in this space are those that are genuinely curriculum-agnostic and focused on outcome measurement: not just delivering content, but tracking whether the training actually results in improved labor market outcomes for participants.
The education platforms that have already achieved meaningful scale — Coursera, which went public in 2021 and has developed enterprise relationships with more than 1,000 organizations globally, and Duolingo, whose market capitalization has reflected strong growth in language learning subscriptions that translate directly into workforce mobility — represent the infrastructure that workforce transitions require. The next generation of workforce development platforms will need to be even more precisely calibrated to the specific skill transitions that AI automation is driving, rather than offering general education that may or may not be relevant to near-term employment.
New Roles, New Economics: What AI Creates
The displacement narrative is only half of the story. AI is also creating new categories of work that did not exist five years ago and that represent significant labor demand. Prompt engineering, AI output evaluation, model fine-tuning, and AI governance are all emerging professional categories with genuine demand that is growing faster than the available supply of qualified workers. More broadly, the organizations deploying AI need professionals who can bridge the gap between AI capabilities and operational implementation — a hybrid role that requires both technical AI literacy and deep domain expertise in the business context where the AI is being deployed.
These new roles share a structural characteristic: they require the combination of technical and domain knowledge that makes them difficult to fill through either traditional technical hiring or traditional domain expert hiring. The legal professional who can effectively supervise and extend Harvey's capabilities is not simply a lawyer who has learned to use a new software tool — they need a genuine understanding of how language models work, what their failure modes are, and how to evaluate their outputs critically rather than trusting them uncritically. That combination is genuinely scarce and will command premium compensation as AI deployment scales.
For investors, this scarcity creates opportunities both in the platforms that help develop AI-hybrid professionals and in the companies that can attract and retain them. Companies with the culture, compensation, and mission alignment to attract AI-native professionals who could command high wages in any organization will have a structural talent advantage over competitors who treat AI deployment as an IT function rather than a strategic capability.
Implications for Capital Allocation
The AI labor market transformation has direct implications for how we evaluate and underwrite investment opportunities at BeMoreeDriven Capital. We have developed a framework for thinking about the labor market exposure of every company we evaluate — both the potential for AI to improve their own productivity and cost structure, and the exposure of their business model to the displacement of their customers' or partners' labor.
Companies whose value proposition depends on providing access to expensive human expertise — at rates that AI can now dramatically undercut — face structural headwinds that need to be explicitly analyzed and addressed in any investment thesis. Companies whose value proposition is amplified by AI — because AI makes their product more powerful, or because their business model benefits from the democratization of expertise — face structural tailwinds that represent a genuine investment edge.
The most interesting opportunities, in our view, are those that sit at the intersection of AI capability and domain expertise depth: companies like Harvey, whose defensibility comes not just from AI capability but from the domain-specific fine-tuning and the legal workflow integration that makes the AI genuinely useful for professional legal work. The general-purpose AI layer will be commoditized — the foundation models from OpenAI, Anthropic, Google, and Meta will compete aggressively on capability and price, with no sustainable moat. The moats will be built in the domain-specific layers: the proprietary training data, the workflow integrations, the trust relationships with professional users, and the accountability infrastructure that makes AI-assisted professional judgment defensible.
The AI labor market transformation is not a story that resolves in a quarter or a year. It is a multi-decade restructuring of the economics of expertise that will create enormous value for the companies that capture the transition intelligently, and enormous disruption for those that are on the wrong side of it. Our goal is to identify and partner with the companies that are building the infrastructure, the applications, and the new professional models that will define what work looks like on the other side of this transition. That work has only begun.