Bell's statistical observation in 1973 was straightforward: the proportion of the workforce engaged in services had crossed the proportion engaged in goods production, and the divergence was accelerating. The conceptual observation was more complex: services were not a single category but a stratified set of activities ranging from personal services (domestic, retail) through professional services (medical, legal, financial) to what Bell called "quaternary" services focused on information and knowledge work. The AI transition operates unevenly across this stratification. It commodifies the professional and quaternary services most directly, leaves the personal services relatively untouched, and creates entirely new service categories organized around human-AI collaboration. The question of who the services serve — and who captures the value they create — is the political question Bell's framework makes unavoidable.
There is a parallel reading that begins not with Bell's employment categories but with the material substrate that makes AI-powered services possible. Every chatbot interaction, every automated document review, every AI-generated image depends on data centers consuming electricity at the scale of small nations, rare earth minerals mined under conditions that recall nineteenth-century extractive colonialism, and cooling water diverted from agricultural regions already facing climate stress. The services economy that Bell analyzed was already an abstraction from material production; the AI services economy is an abstraction squared, hiding its environmental costs behind APIs and cloud interfaces.
The political economy of this shift matters because it determines not just who wins and loses in employment terms, but which regions of the world bear the ecological burden of the transition. The global South provides the lithium for the batteries, the cobalt for the circuits, the low-wage labor for data annotation, while the global North captures the value through platform ownership and high-wage AI engineering jobs. This is not a services economy transitioning through technological change; it is a new form of imperial relationship mediated by computation. The workers displaced from professional services in wealthy countries may suffer wage compression, but the workers mining minerals in Congo or labeling images in Kenya face something closer to pure extraction. Bell's stratification framework captures differences within advanced economies while obscuring the global stratification that makes those differences possible.
Bell's stratification of services has aged well in some respects and poorly in others. The distinction between personal services (face-to-face, low wage, hard to automate) and professional services (credentialed, high wage, now being automated) remains useful. The distinction between services as direct transactions and services as knowledge work has become more important as digital platforms have expanded the second category dramatically. What Bell did not fully anticipate was the emergence of a third layer: services that exist only because of AI — prompt engineering, AI integration consulting, model evaluation — which are simultaneously new job categories and forms of work whose stability depends on the continued scarcity of AI expertise.
The distributional consequences of the services transition have been central to debates about inequality. The professional services that Bell's knowledge class dominated paid premium wages and conferred social status. The personal services that remained stubbornly resistant to productivity gains paid poverty wages and were stigmatized as unskilled. The gap between the two tiers widened throughout the late twentieth century, producing what Bell and his successors called the bifurcated services economy. The AI transition threatens to collapse the professional services into the personal services — not by making the professional services as physical as the personal ones, but by reducing their wage premium toward the level of service work generally.
The question Bell's framework poses — services serving whom? — cuts deeper than the question of who pays for the services. It asks what the services produce, for whom the products are useful, and whose interests are served by the arrangements that organize service delivery. The AI-augmented services that are proliferating in 2025 and 2026 — from customer service chatbots to automated legal review to AI-generated journalism — serve the interests of platform owners and corporate buyers more reliably than they serve the interests of the users who interact with them or the workers who used to produce them. The framework thus directs attention to questions that the celebratory discourse about AI productivity gains systematically avoids.
The policy implications follow directly. If services are stratified and the AI transition affects strata differently, then policy responses must be stratified correspondingly. Uniform responses — retrain workers, expand unemployment insurance, implement universal basic income — address the aggregate effect while missing the structural pattern. The distribution problem that The Orange Pill raises is sharpest in services because services are where the stratification is sharpest, and where the winners and losers are most geographically and occupationally concentrated.
Bell's services analysis drew on the earlier work of Colin Clark and Jean Fourastié on sectoral employment transitions, but Bell extended the framework by disaggregating services into strata with different economic and social dynamics. The analysis appeared most fully in The Coming of Post-Industrial Society and was updated in Bell's subsequent writing as the services sector continued to expand.
Services are stratified, not uniform. Personal, professional, and quaternary services operate by different economic logics and face different automation pressures.
AI automates unevenly across strata. Professional and quaternary services are most exposed; personal services are relatively protected.
New service categories emerge. AI creates its own service layer — prompt engineering, model evaluation, integration consulting — that exists only because AI exists.
Distribution follows stratification. The winners and losers of the AI services transition are concentrated by occupation and geography in ways that aggregate policy responses cannot address.
Whether services can absorb the displaced knowledge workers is contested. Optimists point to the historical expansion of service employment through previous technology transitions. Pessimists argue that the current transition differs because the services most likely to expand (personal, high-touch) pay poorly and lack the autonomy that the displaced knowledge workers valued. Structuralists argue that the answer depends on political choices about wage floors, benefits, and the organization of service work — choices that the aggregate framing of the question systematically obscures.
The right frame depends entirely on the scale at which we examine the services transition. At the occupational level within advanced economies, Edo's reading through Bell is essentially correct (90%): services are indeed stratified, AI does affect strata differently, and the policy implications do follow from understanding these differences. The framework captures the lived experience of knowledge workers watching their expertise get commodified and personal service workers finding their labor suddenly more valuable because it resists automation.
At the global scale, however, the contrarian view dominates (75%): the material and ecological costs of AI infrastructure create new forms of extraction that Bell's framework cannot see. When we ask who provides the compute, who mines the minerals, who labels the training data, we find a different stratification — not between service types but between regions that extract and regions that consume. The question isn't just whether professional services collapse into personal services, but whether entire nations get locked into providing the material substrate for other nations' service economies.
The synthesis requires holding both scales simultaneously. The services economy is transforming both within national boundaries (where Bell's framework applies) and across them (where dependency theory might be more appropriate). The AI transition is simultaneously a story about occupational change in Silicon Valley and mineral extraction in Africa, about prompt engineers in London and data annotators in Manila. The right policy response must therefore operate at both scales: addressing domestic stratification while acknowledging global extraction. This isn't a matter of choosing between frameworks but of recognizing that the services economy has always been embedded in material relations that services discourse tends to obscure.