Engineering the Social Enterprise: Enterprise architecture’s founding paradox

The paradox, stated properly

Enterprise architecture has a well-known origin problem. It is a business management discipline built largely by technologists. It emerged from industrial engineering, computer-integrated manufacturing and software design rather than from management science or organisational sociology, and it inherited the assumptions of those fields: that a system can be decomposed into modules, that interfaces can be specified, that behaviour follows from design.

Applied to an enterprise, this produces a particular picture. The organisation becomes a set of capabilities with defined boundaries, realised by processes, supported by applications, running on infrastructure. Change happens through planned intervention. Coherence comes from governance and standards. It is a clean picture and it is genuinely useful.

It is also, as a description of how organisations actually behave, incomplete in a specific way. Real firms run on informal networks, tacit knowledge, negotiated authority and political bargaining. Decisions get made in corridors and unmade in steering committees. The same nominal capability is enacted differently in Frankfurt and Manila because the people are different. Work gets done through relationships that appear on no diagram.

So far, so familiar — this is the standard framing of EA’s paradox, and it is broadly right. But it understates the problem in one direction and overstates it in another, and both corrections matter for what follows.

This problem is seventy-five years old

The socio-technical critique of engineered work design is not a recent insight EA is catching up to. It has a specific origin, and it is worth knowing because it is more damning than the abstract version.

In 1951, Eric Trist and Ken Bamforth of the Tavistock Institute published a study of the Durham coalfields.¹ Mechanisation had arrived in the form of the longwall method — a technically superior approach that raised theoretical output substantially. It also broke up the small, self-selecting, multi-skilled work groups that had organised themselves at the coalface, replacing them with large shifts of specialised single-task workers. Productivity did not rise as predicted. Absenteeism climbed, accidents rose, and output disappointed. The technical system had been optimised in isolation from the social system it depended on, and the social system took its revenge.

The lesson was not that technology is bad, or that people resist change. It was that the technical and social systems are jointly optimised or not optimised at all.

Two more references belong here, and neither appears in most EA discussions of this topic.

The four-component model that circulates in socio-technical analyses of EA misalignment — structure, task, actor, technology — is Harold Leavitt’s diamond, published in 1965.² It is often cited in recent papers as though it were a recent finding. It is sixty years old, and knowing that changes how one reads the claim: the observation that these four elements must move together is not an emerging discovery about digital transformation, it is a long-established result that architecture practice has repeatedly failed to internalise.

And in 1968, Melvin Conway observed that organisations produce designs that mirror their own communication structures.³ Conway’s Law is usually quoted as a warning about system modularity, but its deeper implication is a direct inversion of the architectural instinct. If the causal arrow runs from social structure to technical structure, then an architect who designs a target architecture without changing the organisation is designing something the organisation cannot build. The org chart is not a constraint on the architecture; it is a substantial part of its specification.

Putting the three together produces a sharper statement of EA’s paradox than the usual one. The problem is not that architects forgot organisations are social. It is that the discipline keeps rediscovering a well-established body of evidence and then proceeding as though the evidence were advisory.

How the engineering instinct shows up in practice

The drift toward formalism

John Zachman’s 1987 framework established that organisations needed a structured way to describe themselves as integrated wholes rather than as piles of isolated technology.⁴ TOGAF, ArchiMate and BIZBOK extended the idea into full ontologies. Each iteration added concepts, relationship types and semantic precision, motivated by an entirely reasonable desire: eliminate ambiguity, enable analysis, support automated validation.

The cost is well described by practitioners who have lived it. The instrument meant to reduce complexity acquires complexity of its own. A statement as simple as “System A uses System B” can require a chain of modelling constructs — precise, and harder to read than the sentence. Meetings drift from the business question to the meta-model question: is this a capability or a service? Should the interface be modelled explicitly? The language overtakes the conversation.

Experienced architects resolve this pragmatically, by adopting a working dialect: a deliberately restricted subset of the notation, models built to answer one question rather than to describe everything. The test is whether a reader says “now I understand how this works” rather than “now I understand the meta-model.” This is sound practice and it is worth naming as such, because it is usually treated as a compromise rather than as the correct method.

The documentation trap

The most expensive failure mode is the attempt to comprehensively catalogue the current state before defining a target. The reasoning is drawn straight from software engineering: understand the system, then change it. But the enterprise keeps moving while you document it, so the model ages faster than it can be completed. Value realisation slides to the right, business leaders conclude that architecture slows things down, and they are not entirely wrong.

The business–IT partition

Conventional frameworks separate business architecture from IT architecture, with technology positioned as a support function executing pre-defined business requirements. That partition made sense when IT ran the back office. It makes considerably less sense when the product is software, the channel is an app, and the operating model is a set of APIs. Separating the layers obscures the feedback loops that now define the business model — which is why business capability maps drawn in isolation from technical reality tend to describe a company that does not exist.

Why EA keeps struggling — and how strong the evidence actually is

It is worth being careful here, because this is an area where confident numbers circulate with weak provenance.

The most solid evidence for EA’s adoption difficulty comes from Forrester’s Modern Technology Operations Survey: enterprise architecture groups exist in around 45% of organisations globally, while 30% report that their company had such a department and disbanded it. Forrester’s own reading is that teams get established, fall into recognisable traps — the “ivory tower” being the archetype — get disbanded, and are re-formed later in a repeating cycle.⁵ The form–disband–reform loop is a more interesting finding than any single satisfaction statistic, because it suggests the function meets a real need in a way that current practice does not durably satisfy.

A widely repeated claim that 32% of EA deliverables are never reused is often attributed to Gartner. It traces, as far as I can establish, to an EA tool vendor’s blog rather than to a locatable Gartner publication — and the vendor’s own gloss describes 32% as meaning “the majority of what EA teams produce is being quietly ignored,” which is not what 32% means.⁶ The underlying point about unused artefacts is credible from experience. The number should not be cited as though it were measured.

Beneath the statistics, the recurring failure patterns are consistent enough to name. Four dimensions of misalignment show up repeatedly:

Four recurring socio-technical misalignments. Each is a mismatch between a formal model and an informal reality — not a failure of modelling rigour.

There is a harder position worth acknowledging rather than dismissing: that EA is frequently low-value ceremony, and that the disband rate reflects accurate assessment rather than organisational short-sightedness. The honest response is that both readings fit the data. Architecture teams that produce artefacts nobody uses are correctly disbanded. That some organisations then re-form them suggests the underlying coordination problem is real. Which of these describes any particular case is an empirical question about that case, not something the discipline can settle in its own favour by assertion.

Culture is the variable, not the backdrop

If the social system determines whether architecture takes, then culture is not a soft consideration to acknowledge in a closing paragraph. It is the primary design constraint.

The Competing Values Framework — developed by Quinn and Rohrbaugh in the early 1980s and popularised in Cameron and Quinn’s later work — maps culture on two axes, flexibility versus control and internal versus external focus.⁷ It is a serviceable diagnostic for architects, because it predicts which architectural posture will be accepted:

The same architecture practice succeeds or fails depending on cultural fit. This is a diagnostic aid, not a taxonomy of organisations — most firms contain several of these at once.

Two further cultural variables deserve specific mention because they bear directly on model accuracy.

Psychological safety. Amy Edmondson’s work established that teams differ substantially in whether members can raise problems without fear of penalty.⁸ For architecture this is not a wellbeing concern, it is a data quality concern. Where safety is low, status reports get sanitised, technical debt goes unreported, and the architecture repository fills with fiction. You cannot model a system you are not being told the truth about.

Power distance. In organisations with steep authority gradients, junior engineers do not challenge senior designs, so flawed designs survive review.⁹ Architecture review is only a control if dissent is possible in the room.

And organisations are rarely culturally uniform. The business/IT subcultural divide — one side valuing speed and revenue, the other stability and standardisation — remains a primary source of architecture failure. The architect’s job in that gap is translation, and translation is a skill, not a disposition.

What actually seems to work1. The Historical Pivot: Why Socialism Attached Itself to Labor

Socialism’s historical alignment with the industrial working class reflected the material conditions of the First Industrial Revolution rather than an arbitrary ideological choice. Between roughly 1750 and 1850, the shift from agriculture and artisanal craft to mechanized factory production reorganized the British and, subsequently, the wider European economy.¹ Land and capital concentrated in the hands of a rising industrial bourgeoisie, and large numbers of small-scale producers lost their independent means of subsistence and entered urban wage labor.

In this phase, labor was the binding constraint on production. Early steam engines, spinning jennies and blast furnaces multiplied output, but they were crude machines requiring continuous human coordination, tending and manual power. Production was labor-intensive in a straightforward technical sense: expanding output generally meant employing more workers.

Classical political economy and early socialist thought were both responding to this reality, though they interpreted it differently. David Ricardo, in the chapter “On Machinery” added to the third edition of his Principles in 1821, conceded that mechanization could be injurious to the interests of the laboring class — a notable admission from a writer otherwise confident about capital accumulation.² Economic historians have since documented a prolonged period, which Robert Allen labelled “Engels’ pause,” during which British output per worker rose while real wages stagnated, roughly from the 1790s to the 1840s.³

Marx and Engels built a more comprehensive critique on this foundation. In Marx’s framework, constant capital — machinery, infrastructure and raw materials — transfers its existing value to the final commodity, while only living labor (variable capital) creates new value. Profit, on this account, derives from surplus value: the portion of value produced by the worker beyond the cost of reproducing their labor-power.⁴

It is worth stating plainly that the labor theory of value is a contested theoretical position rather than a settled empirical finding. Mainstream economics since the marginalist revolution of the 1870s explains factor incomes through marginal productivity and relative scarcity instead, and the two frameworks generate materially different predictions about what automation does to profit. What follows should be read as a reconstruction of the socialist argument and the evidence bearing on it, not as a demonstration that the framework is correct.

Whatever the correct theory of value, the strategic implication was real and is not in dispute. Because accumulation depended on employing large workforces at fixed physical sites, workers held structural leverage: an organized withdrawal of labor could halt production. That leverage underwrote trade unions, collective bargaining institutions and mass labor parties throughout the twentieth century.

In Germany this trajectory produced the post-war system of Mitbestimmung (codetermination). The Montan-Mitbestimmung of 1951 established genuine parity between shareholder and employee representatives in coal and steel; the broader Codetermination Act of 1976 extended employee representation to roughly half the supervisory board seats in large firms, but reserved a casting vote for the chairman, who is elected by the shareholder side.⁵ Codetermination therefore institutionalized labor as a powerful partner rather than a strictly equal one — a distinction that matters when assessing what the system can and cannot achieve under pressure. For most of a century, the mobilization and protection of the industrial working class was the organizing purpose of socialist politics, and “the worker” was its central political subject.

Table 1. Structural comparison. The right-hand column describes tendencies that are documented but whose causes and permanence remain debated.

2. The Fourth Industrial Revolution and the Labor–Productivity Debate

The claim that concerns contemporary socialist theory is that the link between labor input and economic output has weakened, and that this weakening is structural rather than cyclical. The evidence is real, but it is more contested than it is usually presented.

Under a standard neoclassical account, productivity growth raises the marginal product of labor and, in competitive markets, feeds through to wages. Since roughly 1980, that relationship has looked weaker across most advanced economies. The labor share of national income — compensation as a proportion of value added — has declined across a wide set of countries and industries, most sharply during the 2000s.

Two qualifications matter, and the original framing of this argument tends to omit both.

First, the magnitude of the decline depends on disputed measurement choices. Rognlie (2015) finds that the rise in the net capital share is driven almost entirely by housing.⁶ Bridgman argues the picture changes substantially once depreciation is netted out. Koh, Santaeulàlia-Llopis and Zheng (Econometrica, 2020) argue that much of the measured U.S. decline is an artifact of the Bureau of Economic Analysis reclassifying intellectual property spending from expense to investment.⁷ Most economists nonetheless accept that a real decline has occurred; its size and causes remain open questions.

Second, there is no settled verdict on the cause. Candidate explanations include automation and capital–labor substitution, but also globalization and offshoring, the rise of high-markup “superstar” firms (Autor et al., 2020), declining union density, and weakening product-market competition.⁸ Attributing the decline primarily to technology is a defensible position, not an established finding.

What the automation evidence actually shows

Acemoglu and Restrepo’s distinction between displacing and reinstating technologies is the most useful analytical frame here: whether automation reduces labor demand depends on whether it merely substitutes for existing tasks or also generates new tasks in which labor holds a comparative advantage.⁹ Their concern is that the recent technology mix has been unusually displacement-heavy — including “so-so automation” that eliminates jobs without delivering large productivity gains. This is an empirical claim about the present, not a law of technological development.

A study of European regions estimates that a doubling of regional AI patenting is associated with a decline of roughly 0.7 to 1.6 percentage points in the local labor share.¹⁰ This is a useful finding, but it is a regional association reported in a working paper, not proof of a mechanical transfer of wealth from labor to capital. Regions that innovate intensively differ from other regions along many dimensions, and the paper should be cited with that caveat attached.

The distributional argument is considerably more robust than the causal one. Because capital ownership is far more concentrated than labor income, any shift of factor income from labor to capital tends, other things equal, to widen wealth inequality. That inference does not depend on resolving the debate about causes.

Sectoral figures, with their provenance

Several figures circulate widely in this literature and are worth restating with their sources and dates attached, because a number of them are older than their usage implies:

●        Heavy industry. U.S. steel output per employee is several times higher than in the early 1980s. Much of this reflects the shift from integrated mills to electric arc furnaces and the closure of the least productive capacity, alongside automation — not automation alone.¹¹

●        Retail and commerce. The Council on Foreign Relations’ The Work Ahead task force report documented net losses in brick-and-mortar retail employment alongside smaller gains in e-commerce warehousing. That report was published in 2018; U.S. warehousing employment has grown substantially since, so the net position is now contested and the figures should not be presented as current.¹¹

●        Low-wage work. The White House Council of Economic Advisers reported in 2016 that 83 percent of jobs paying under $20 an hour — in 2016 dollars — “would come under pressure from automation.”¹² The estimate was produced by ranking occupations against Frey and Osborne’s occupation-level automation probabilities,¹³ a methodology criticized for treating whole occupations rather than constituent tasks as automatable. The OECD’s task-based replication produced far lower estimates, on the order of 9 percent of jobs at high risk.¹⁴ Both numbers deserve to appear together.

●        Cognitive and professional work. Demand for advanced digital skills has grown rapidly while wages for routine cognitive work in finance, journalism and law have come under pressure. The specific “quadrupled to 23 percent of employment” figure also originates in the 2018 CFR report and should be dated accordingly.¹¹

Historically, the labor share remained comparatively stable because technological change was broadly balanced: automation eliminated tasks while new tasks emerged elsewhere. Whether generative AI breaks that pattern is genuinely unknown. The Federal Reserve Bank of Philadelphia treats it as a plausible turning point;¹⁵ others expect the historical pattern of task creation to reassert itself. The defensible position is that the question is open — which also means it should not be assumed in the direction of concluding that traditional labor politics has been rendered obsolete.

Table 2. Frequently cited figures and what they support. Each is defensible when stated with its source and date; none supports a stronger causal claim than its method allows.

3. Volkswagen and the German Automotive Crisis

The strain on the German model of social partnership is visible in the current crisis at Volkswagen. The situation is unresolved, and it is important not to describe a contested proposal as a settled outcome.

For decades the German automotive sector was the institutional center of that model: high-wage long-term employment, sector-wide collective bargaining, and a codetermination system giving IG Metall and works councils substantial influence over strategic planning.¹⁶ The transition to battery electric vehicles and software-defined platforms, combined with competition from vertically integrated Chinese manufacturers, U.S. import tariffs and high domestic energy costs, has placed that arrangement under severe pressure.

Status of the restructuring

In June and July 2026, reporting based on leaked documents indicated that CEO Oliver Blume would present a plan — a “Group Target Picture 2030” — that could reduce headcount by up to 100,000 and wind down production at four German sites: Emden, Zwickau, the Hanover commercial vehicle plant, and Audi’s Neckarsulm facility.¹⁷ The 100,000 figure is a cumulative total. It would roughly double the approximately 50,000 reduction already agreed with unions in late 2024, of which more than 28,000 departure agreements had reportedly been signed by the June 2026 annual general meeting.

The plan has not been adopted. At the supervisory board meeting on 9 July 2026, labor representatives blocked the proposals. Volkswagen subsequently announced narrower measures — reducing the model range by up to half and cutting capacity to around nine million vehicles a year — without confirming the larger job cuts.¹⁸ In an internal memo the following week, Blume described the 100,000 figure as a theoretical calculation derived from an estimated 20 percent cost gap against comparable rivals, stated that a competitive use for the four named plants in the 2030s could not currently be confirmed, and expressed a preference for what he called “intelligent solutions” over outright closure.¹⁹ Any account presenting the cuts or the closures as decided is premature.

The financial pressure behind the proposal is not in dispute: first-quarter 2026 net profit fell 28 percent year on year to €1.56 billion, U.S. tariffs are estimated to cost the group roughly €4 billion annually, and first-quarter sales in China — the group’s largest single market — fell about 20 percent.

Overcapacity and asset disposals

European assembly capacity substantially exceeds demand. Industry estimates commonly put the gap at several million vehicles a year, but such figures are sensitive to assumptions about shift patterns and utilization thresholds and are best treated as approximations rather than precise measurements. Data cited by Reuters put Volkswagen’s German plants at about 81 percent of standard capacity in 2026, falling to roughly 73 percent by 2030, with Zwickau projected to decline from 88 percent to 42 percent over the same period.¹⁷

Volkswagen has also monetized assets. In June 2026 it agreed to sell 51 percent of Everllence — the large marine and power-generation engine business formerly known as MAN Energy Solutions — to Bain Capital, in a leveraged transaction generating approximately €7.4 billion in proceeds; Volkswagen retains 49 percent.²⁰ The €7.4 billion figure represents total proceeds, including debt raised at completion and revaluation, rather than a purchase price for the stake alone. Whether such disposals are best described as prudent balance-sheet management or as financialized asset stripping is a matter of interpretation, and the same transaction supports both readings.

The “VW Act”

The Volkswagen Act, enacted on 28 July 1960, does not grant Lower Saxony its shareholding. The state holds roughly 20 percent of voting rights independently; the statute raises the threshold for major shareholder resolutions above 80 percent, which gives any holder of 20 percent an effective blocking minority. A separate provision requires a two-thirds supervisory board majority to establish or relocate production sites. The European Court of Justice struck down parts of the law in 2007; Germany amended it in 2008 while retaining the supermajority requirement, and in 2013 the Court accepted the amended version.²¹ Notably, the statute does not mention plant closures explicitly, so whether it constrains closure as distinct from relocation is legally contested — a point worth registering, since the law’s protective reach is central to the current dispute.

The electric vehicle labor paradox

The “fewer moving parts” claim is frequently garbled in circulation. The familiar comparison — roughly 20 moving parts against 2,000 — applies to the powertrain, not to the vehicle as a whole; total vehicle part counts, on the order of 30,000, differ far less between drivetrains. From this, automakers and analysts have projected that battery electric vehicle assembly would require around 30 percent less labor.²²

Recent evidence complicates that projection. Weng, Ahmed, Ehrlich and Stefanopoulou (Nature Communications, 2024) examined U.S. plants that fully transitioned to BEV assembly and found that labor intensity rose.²³ At the Alameda County site — the former NUMMI plant now operated by Tesla — labor intensity reached its most efficient recorded level of about 15 workers per 1,000 vehicles per year (WPV) in 2006 under combustion-vehicle production. Under Tesla’s BEV production it averaged roughly 51 WPV between 2019 and 2022, rising to about 67 WPV once battery cell and pack workers at Sparks, Nevada are included.

Two caveats are essential, and the study itself supplies both. First, the comparison is confounded by vertical integration: Tesla performs in-house a range of work that NUMMI outsourced. Second — and this is the authors’ own headline conclusion — they interpret their findings as evidence that rapid, widespread job loss at vehicle assembly plants is a smaller risk than many fear, and they estimate that BEV plants may take more than fifteen years to reach labor-intensity parity with combustion plants. The study therefore does not establish that electric vehicles are permanently more labor-intensive. It establishes that the transition is slower and more labor-hungry than the 30 percent figure implied, which is a different and more limited claim.

The more substantive competitive issue for German manufacturers is where the labor-dense, high-value work sits. Battery cells, cathodes and anodes are produced largely abroad, principally in China. To compete on cost, European producers must either localize that segment of the value chain or automate domestic assembly further.²² Both management and union accounts of the “fewer parts” argument should be read with an awareness of interest: the claim serves management as a rationale for headcount reduction, and unions have a corresponding interest in disputing it.

The union response

IG Metall mobilized workers at around twenty Volkswagen sites ahead of the July board meeting; IG Metall President Christiane Benner and General Works Council Chair Daniela Cavallo publicly opposed the plans, and labor representatives used their supervisory board position to block them.¹⁷²⁴

How to characterize that response is a question of political judgment rather than fact, and both available readings should be stated. Critics on the left — notably the World Socialist Web Site, a publication of the Socialist Equality Party whose coverage should be read as advocacy rather than as neutral reporting — argue that the union leadership has functioned as a stabilizing force, prioritizing institutional defense of codetermination and the VW Act over the defense of jobs, having already accepted substantial concessions in 2024.²⁵ Others read the same sequence as an effective use of the institutional leverage codetermination provides: the proposal was, after all, blocked. The documentary record supports both interpretations, and a reader is entitled to weigh them.

Reporting also indicates that Volkswagen has for some time sought a defense-sector partner for its underutilized Osnabrück plant, as one of several alternative uses under consideration.¹⁷ More specific claims circulating about the identity of prospective partners and about internal disciplinary practices derive from partisan sources and are not independently corroborated; they have been omitted here rather than repeated.

Table 3. Volkswagen restructuring: agreed positions, proposals, and confirmed outcomes as at late July 2026. Figures reported in the press from leaked documents are identified as proposals throughout.

4. The Theoretical Dilemma: Socialism Without the Proletariat

Suppose, for the sake of argument, that labor’s structural weight does continue to decline. This would pose a genuine problem for a tradition that located political agency at the point of production.

The socialist movement’s historical power derived from the worker’s position within an indispensable production process. If a factory, logistics network or software platform can operate with substantially fewer people, the threat of withholding labor loses force. On this reading, a population displaced from production loses the leverage needed to bargain over distribution, and its political agency is reduced to dependence on state transfers. Whether this describes a realistic trajectory or an overextrapolation from a decade of data is precisely what is at issue.

Contemporary theorists have offered several responses.

Post-work and fully automated luxury communism

Aaron Bastani, and separately Nick Srnicek and Alex Williams, argue that converging advances in artificial intelligence, robotics, synthetic biology and renewable energy could substantially reduce material scarcity.²⁶²⁷ On this view, automation is not a catastrophe to be resisted but a precondition of liberation. Decoupling income from work — typically through a universal basic income — would allow necessary labor to be performed largely by machines while human activity is freed for self-development and creative work.

Critics on the left, notably in Catalyst, raise three objections.²⁸ The first concerns necessary labor: for a basic income to confer real freedom it must retain purchasing power, which requires a stable supply of essential goods and services. If agriculture, infrastructure maintenance, logistics and care work are not fully automated — and there are strong engineering reasons to doubt they will be soon — society remains dependent on human labor in exactly the sectors that are hardest and least desirable. Making that work voluntary risks shortages and inflation, eroding the transfer’s value. The second objection is that cash distribution mimics market logic, treating citizens as consumers rather than as participants in economic planning. The third is political: some free-market advocates support basic income precisely as a replacement for public services, in which case it functions as a mechanism for privatization rather than a bridge beyond capitalism.²⁹

These are serious objections, though not decisive ones. Basic income pilots have generally found modest rather than dramatic reductions in labor supply, and proponents note that most designs are intended to supplement public services rather than replace them. The disagreement is partly empirical and remains unresolved.

Techno-feudalism

Yanis Varoufakis argues that large technology platforms have moved beyond conventional capitalist competition: rather than maximizing profit through the employment of wage labor, they enclose digital infrastructure and extract “cloud rent” from consumers and from other firms obliged to use their platforms.³⁰ If rent extraction has displaced surplus-value production, the traditional socialist focus on workplace organization would be substantially bypassed.

The thesis is influential but far from accepted. Evgeny Morozov and a number of Marxist economists argue that platform rents remain internal to capitalism rather than superseding it, that profit-seeking through wage labor remains the dominant mode across the wider economy, and that the feudal analogy obscures more than it clarifies.³¹ Readers should treat “techno-feudalism” as one interpretive framework among several, and its vocabulary — “digital lords,” “cloud serfs” — as rhetorical rather than descriptive.

5. Responses and Proposals

If the diagnosis is uncertain, the prescriptions are more uncertain still. Rather than advocating one program, this section sets out the principal families of response and the main considerations for and against each.

●        Adjustment through education and the existing safety net. The market-liberal position holds that labor markets have absorbed comparable transitions before and will again, and that policy should focus on retraining, wage insurance and portable benefits. For: the historical record of task creation is genuinely strong. Against: adjustment costs have fallen heavily on particular regions and cohorts, and retraining programs have a mixed evaluation record.

●        Preservationist industrial policy. Subsidy, tariff protection and negotiated employment guarantees to sustain existing industrial employment. For: it protects communities during transition and preserves productive capability. Against: it can entrench uncompetitive cost structures, as the European automotive dispute illustrates, and postpones rather than resolves adjustment.

●        Post-work and basic income. Decoupling income from employment. For: it addresses distribution directly and does not depend on preserving particular jobs. Against: the necessary-labor and fiscal objections above remain unresolved, and the political coalition supporting it is unstable across the left–right divide.

●        Universal basic services. Expanding in-kind provision — housing, transport, healthcare, education, and potentially digital infrastructure — rather than cash transfers.³² For: it insulates essential needs from the price mechanism and avoids the inflation objection to basic income. Against: it requires substantial state administrative capacity and constrains individual choice; costings are contested.

●        Working-time reduction and shared labor. Distributing necessary work more evenly and shortening the standard week. For: four-day-week trials in several countries have reported broadly positive outcomes for wellbeing with limited productivity loss. Against: trial evidence is short-term, largely from white-collar settings, and does not straightforwardly generalize to sectors with capacity constraints.

●        Broadening capital ownership. Employee share ownership, sovereign wealth funds, citizens’ capital funds, and Varoufakis’s “one employee, one share” proposal.³⁰ For: it addresses the distributional problem at its source, since concentrated capital ownership is what converts a falling labor share into rising inequality. Against: the transition path from current ownership structures is politically formidable, and the governance record of large collectively owned enterprises is mixed.

These options are not mutually exclusive, and the most plausible programs combine several. What can be said with reasonable confidence is narrower than the strongest claims on either side: the labor share has fallen; concentrated capital ownership converts that fall into rising inequality; and the causes are multiple, with automation one among several. Policy responses that address distribution and ownership directly are therefore less dependent on resolving the causal debate than responses premised on a particular theory of technological change.

Conclusion

The relationship between labor input and economic output has changed since the mid-twentieth century, and the labor share of national income has declined across most advanced economies. This much is well documented, even if its magnitude and causes remain subject to active dispute among economists. The classical socialist strategy of relying on the industrial workforce as the primary source of political leverage has correspondingly weakened, though “weakened” is a more defensible description than “obsolete.”

The Volkswagen dispute illustrates the strain on codetermination under competitive pressure, but it also illustrates the limits of confident prediction. Labor representatives blocked the restructuring proposal in July 2026, which is not the behavior of an institution without leverage. Whether that constitutes an effective defense or a delaying action will only be clear in retrospect.

Similarly, the evidence on electric vehicle assembly cuts against the simplest displacement narrative: the best available study finds that transitioned plants have so far required more labor, not less, and its authors read this as reason for less alarm rather than more. Where the evidence is genuinely mixed, it is more useful to say so than to select the reading that fits a prior commitment.

What follows for policy is that the distributional question is more tractable than the technological one. Regardless of how fast automation proceeds, the concentration of capital ownership determines whether productivity gains are broadly or narrowly shared. Proposals that address ownership, working time and the provision of essential services can be evaluated on their own merits, without first having to settle whether artificial intelligence represents a genuine break in economic history or another chapter in a long and familiar pattern.

Note on sources

The reference list below differs from the original in three respects, in the interests of verifiability:

●        Discussion-forum threads, personal blog posts and reader-review pages have been replaced with the primary literature they were standing in for — for example, Marx and Ricardo directly rather than secondary summaries, and the peer-reviewed labor-share literature rather than commentary about it.

●        The Volkswagen narrative is now sourced primarily to wire and mainstream business reporting. The World Socialist Web Site is retained where it is the origin of a specific interpretive claim, and is identified in the text as a partisan publication so that readers can weigh it accordingly.

●        Several statistics have been re-anchored to their originating documents — the 2016 Council of Economic Advisers report and the 2018 Council on Foreign Relations task force report — with their publication dates stated, since both were being cited as though current.

Works cited

1. The First Industrial Revolution: Creation of a New Global Human Era. MPRA Paper 96644. https://mpra.ub.uni-muenchen.de/96644/1/MPRA_paper_96644.pdf

2. Ricardo, D. (1821). On the Principles of Political Economy and Taxation, 3rd ed., ch. 31, “On Machinery.” See also Acemoglu, D. (2024), Learning from Ricardo and Thompson: Machinery and Labor in the Early Industrial Revolution and in the Age of Artificial Intelligence, MIT Economics. https://economics.mit.edu/sites/default/files/2024-10/Learning%20from%20Ricardo%20and%20Thompson%20-%20Machinery%20and%20Labor%20in%20the%20Early%20Industrial%20Revolution%20-%20and%20in%20the%20Age%20of%20AI.pdf

3. Allen, R. C. (2009). Engels’ pause: Technical change, capital accumulation, and inequality in the British industrial revolution. Explorations in Economic History, 46(4), 418–435.

4. Marx, K. (1867/1976). Capital: A Critique of Political Economy, Volume I. Penguin Classics. See esp. Parts III–IV on surplus value and constant capital.

5. Klikauer, T. 50 Years of Co-Determination in Germany. https://www.filmsforaction.org/articles/50-years-of-codetermination-in-germany/

6. Rognlie, M. (2015). Deciphering the Fall and Rise in the Net Capital Share. Brookings Papers on Economic Activity, 2015(1), 1–69. https://www.brookings.edu/wp-content/uploads/2016/07/2015a_rognlie.pdf

7. Koh, D., Santaeulàlia-Llopis, R., & Zheng, Y. (2020). Labor Share Decline and Intellectual Property Products Capital. Econometrica, 88(6), 2609–2628.

8. Autor, D., Dorn, D., Katz, L. F., Patterson, C., & Van Reenen, J. (2020). The Fall of the Labor Share and the Rise of Superstar Firms. Quarterly Journal of Economics, 135(2), 645–709. https://economics.mit.edu/sites/default/files/publications/Autor%20et%20al_2020_The%20Fall%20of%20the%20Labor%20Share%20and%20t.pdf

9. Acemoglu, D., & Restrepo, P. (2019). Automation and New Tasks: How Technology Displaces and Reinstates Labor. Journal of Economic Perspectives, 33(2), 3–30.

10. Unslicing the pie: AI innovation and the labor share in European regions. EconStor working paper. https://www.econstor.eu/bitstream/10419/315088/1/1909805009.pdf

11. Council on Foreign Relations (2018). The Work Ahead: Machines, Skills, and U.S. Leadership in the Twenty-First Century — Findings. https://www.cfr.org/task-force-reports/work-ahead/findings

12. Executive Office of the President (2016). Artificial Intelligence, Automation, and the Economy. https://obamawhitehouse.archives.gov/sites/whitehouse.gov/files/documents/Artificial-Intelligence-Automation-Economy.PDF

13. Frey, C. B., & Osborne, M. A. (2017). The future of employment: How susceptible are jobs to computerisation? Technological Forecasting and Social Change, 114, 254–280.

14. Arntz, M., Gregory, T., & Zierahn, U. (2016). The Risk of Automation for Jobs in OECD Countries. OECD Social, Employment and Migration Working Papers No. 189.

15. Federal Reserve Bank of Philadelphia (2024). Generative AI: A Turning Point for Labor’s Share? Economic Insights, Q1 2024. https://www.philadelphiafed.org/-/media/frbp/assets/economy/articles/economic-insights/2024/q1/eiq124-generative-ai-a-turning-point-for-labors-share.pdf

16. Social Partnership in Crises: Transitions and Strategic Participation of Collective Labour in German Automotive Corporations. Review of Radical Political Economics (Taylor & Francis). https://www.tandfonline.com/doi/full/10.1080/07360932.2026.2667295

17. Reuters / Euronext (July 2026). Volkswagen board faces crunch talks over plant closures and job cuts. https://live.euronext.com/en/financial-news/volkswagen-board-faces-crunch-talks-over-plant-closures-and-job-cuts

18. CNBC (10 July 2026). Volkswagen to slash model lineup and shrink capacity — but no word on job cuts. https://www.cnbc.com/2026/07/10/volkswagen-vw-autos-job-cuts-germany.html

19. Quartz (13 July 2026). Volkswagen CEO Oliver Blume warns of 100,000 total job cuts [internal memo]. https://qz.com/volkswagen-ceo-job-cuts-100000-memo-071326

20. Reuters (25 June 2026). Volkswagen shares boosted by Everllence deal promising €7.4 billion in proceeds. https://wmbdradio.com/2026/06/25/volkswagen-shares-boosted-by-everllence-deal-promising-e7-4-billion-in-proceeds/

21. Court of Justice of the European Union, Case C-112/05, Commission v Germany, judgment of 23 October 2007; and the subsequent 2013 judgment on the amended statute. https://curia.europa.eu/jcms/jcms/P_29602/es

22. U.S. International Trade Commission. How Does Increased EV Production Affect U.S. Automotive Employment? https://www.usitc.gov/publications/332/executive_briefings/ebot_how_does_increased_ev_production_affect_employment.pdf

23. Weng, A., Ahmed, O. Y., Ehrlich, G., & Stefanopoulou, A. (2024). Higher labor intensity in US automotive assembly plants after transitioning to electric vehicles. Nature Communications, 15, 8088. https://www.nature.com/articles/s41467-024-52435-x

24. The Guardian (9 July 2026). VW workers protest in Germany over proposed job cuts and factory closures. https://www.theguardian.com/business/2026/jul/09/vw-protests-germany-proposed-job-cuts-factory-closures

25. World Socialist Web Site (27 June 2026). Enough is enough: Defend every job at Germany’s Volkswagen. Publication of the Socialist Equality Party; cited as a partisan source. https://www.wsws.org/en/articles/2026/06/27/vhmg-j27.html

26. Bastani, A. (2019). Fully Automated Luxury Communism: A Manifesto. Verso.

27. Srnicek, N., & Williams, A. (2015). Inventing the Future: Postcapitalism and a World Without Work. Verso.

28. Post-Work Socialism? Catalyst: A Journal of Theory and Strategy (2022). https://catalyst-journal.com/2022/09/post-work-socialism

29. Universal basic income: utopian dream or libertarian nightmare? The Communist. https://communist.red/universal-basic-income-utopian-dream-or-libertarian-nightmare/

30. Varoufakis, Y. (2023). Technofeudalism: What Killed Capitalism. The Bodley Head.

31. Morozov, E. (2022). Critique of Techno-Feudal Reason. New Left Review, 133/134.

32. Coote, A., & Percy, A. (2020). The Case for Universal Basic Services. Polity Press.

The practices that survive contact with organisations share a common property: they trade control for influence, and accept partial coherence as the achievable goal.

●        Guardrails over rules. Define boundaries and principles, then let teams decide inside them. The useful analogy is lane assist rather than braking — continuous, low-friction correction rather than periodic hard stops.

●        Federated governance. Autonomy at the edge, alignment at the core. This works where trust and communities of practice are mature, and degrades into fragmentation where they are not. It is not a universally superior model; it is a model with prerequisites.

●        Architecture Decision Records. Short documents capturing a decision, its context, its rationale and its consequences. Their real value is organisational memory: they prevent the Chesterton’s Fence problem of removing a constraint without knowing why it was put there. They also democratise decision-making by making reasoning inspectable rather than authoritative.

●        A restricted modelling dialect. As above: model to answer a question, not to represent the meta-model.

●        Metrics in the language of the person funding you. Cost avoidance, total cost of ownership per capability, time to business value, portfolio agility. Architecture that reports on standards compliance is reporting on its own activity, not on outcomes.

●        Scenario modelling. Simulating trade-offs — what does retiring this application actually cost, which consolidation path carries the best risk-adjusted return — shifts architecture from documentation to decision support. This is where EA tooling earns its keep, and it is the strongest available answer to the “what are you for” question.

The agentic turn: what genuinely changes

Autonomous agents — software that reasons over a goal, invokes tools, and coordinates with other agents rather than waiting for instructions — do represent a real shift, and it is worth being precise about what the shift is.

Traditional enterprise software relied on human bandwidth to supply context, judge exceptions and trigger transactions. The application was passive; the person was the active element. When an agent executes an onboarding sequence or a reconciliation, that changes: software becomes an operational actor. In ArchiMate terms, agents sit awkwardly across the layer boundary — modelled as Business Roles when they replace or augment human function, and as Application Components when they execute software services. That modelling awkwardness is not a notation defect; it is the notation faithfully reporting that a long-standing distinction has stopped being clean.

Four architectural consequences follow, and these are the substantive parts of the agentic-EA agenda:

●        Latency. Agents reasoning continuously against enterprise state need data far fresher than overnight batch cycles provide. Streaming and event architectures move from optimisation to precondition.

●        Semantics. An agent reasoning over ambiguous data reasons badly. Machine-readable ontologies and knowledge graphs shift from a data-management nicety to an operational dependency, because the model has no colleague to ask what the field actually means.

●        Interface standardisation. Heavily customised legacy systems are difficult to orchestrate. Reducing bespoke modification and exposing standard interfaces is what makes systems addressable by agents. (Several vendors market proprietary programmes under branded names for this; the underlying principle is vendor-neutral and predates the branding.)

●        Lifecycle management. Agents need to be inventoried, given scoped identities, linked to the capabilities they serve, monitored and eventually decommissioned. Without this, organisations accumulate orphaned agents the way they accumulated orphaned servers.

One economic figure deserves correction, because it is widely misattributed. The estimate of $2.6–$4.4 trillion in annual value is McKinsey’s June 2023 assessment of generative AI across 63 use cases in 16 business functions — not an estimate of agentic AI, and not a measurement of realised value.¹⁰ It is a modelled potential, produced before agentic deployment was widespread. Citing it as the prize for agentic architecture attributes a specific number to a different technology.

Why agents sharpen the paradox rather than resolve it

Here is the claim I want to push back on, because it is the most appealing and the least supported idea in the current agentic-EA discussion: that autonomous execution reconciles EA’s founding contradiction. The argument runs that when abstract capability models become executing code, the gap between the engineered blueprint and the messy social organisation finally closes — the model becomes the operation.

It is a satisfying story, and I think it is wrong in four ways.

1. Only the codifiable part becomes executable

Capability models have always described a mix of the formalisable and the non-formalisable. Agents can execute the formalisable portion — which is, by definition, the portion that was already the easiest to specify. What remains is exception handling under ambiguity, negotiation across units with conflicting incentives, judgement calls where the rule under-determines the answer, and the tacit knowledge that never made it into any document. Those are precisely the parts that resisted thirty years of process formalisation. There is no reason to expect them to yield now, and automating around them does not dissolve them. It isolates them.

2. The politics move upward, not away

Encoding a capability into an agent requires settling what the capability is — which definition, whose thresholds, which exception path, optimising for which objective. Those questions were previously left productively ambiguous, and the ambiguity absorbed a great deal of organisational conflict. Codifying them does not remove the disagreement; it forces it to resolution and then freezes the result in executing code, where changing it requires a release rather than a conversation.

An agent is a political settlement with an execution engine attached. Making capability definitions binding raises the stakes of the definition fight rather than ending it.

And agents themselves become political objects. Who owns this agent. Whose budget funds it. Whose KPI does it optimise when two units want different things. Which team is accountable when it acts wrongly. These are not technical questions and they will be settled by the same informal, negotiated, power-inflected processes that settle everything else.

3. Conway’s Law does not switch off

If organisations produce designs mirroring their communication structures, then a fragmented organisation will build a fragmented agent estate. Multi-agent architectures will reproduce existing silo boundaries, because the teams building them sit inside those boundaries. Agentic architecture is not an escape from organisational structure; it is another medium in which organisational structure is expressed — and a faster-moving one.

4. The Tavistock result applies directly

The Durham coalfield study is the closest available historical analogue. A technically superior method was introduced without redesigning the social organisation of work, and performance fell short of projection. Organisations now deploying agents into workflows without redesigning roles, accountability and decision rights are running the same experiment. Deloitte’s 2026 survey finding that 84% of companies have not redesigned roles around AI, while around three-quarters plan agentic deployment within two years, describes exactly that configuration.¹¹

Lisanne Bainbridge’s “ironies of automation” adds the operational corollary.¹² Automating the tractable parts of a process leaves human operators with the residual hard cases, while eroding, through disuse, the situational awareness those cases require. Applied to agentic enterprises: the humans left in the loop will handle only the exceptions, will handle fewer of them, and will therefore be less practised at exactly the moments when their judgement matters most. That is an argument for deliberate role design, not an argument against automation — but it is not an argument that the paradox has dissolved.

So the more defensible position is this. Agentic AI does not resolve the tension between engineered models and social systems. It removes the slack that used to absorb it. When execution was human, a badly specified capability model was survivable because people quietly did the sensible thing instead. Agents do not do the sensible thing instead; they do the specified thing at scale and at speed. The cost of a wrong model goes up, and the interval between error and consequence goes down.

That is a strong argument for taking architecture seriously. It is not an argument that architecture has become easy.

Governance for hybrid human–agent organisations

The governance requirements follow from the above, and they are more concrete than most of what EA has previously been asked to govern.

●        Identity and least privilege. Every agent needs a distinct software identity with scoped, revocable permissions. Agents inheriting a human’s credentials is the current default in many deployments and it is untenable at scale.

●        Pre-execution policy evaluation. A control point that evaluates proposed actions against policy, legal constraint and risk threshold before execution, with automatic escalation to a human above defined risk boundaries.

●        Observability. Capture of reasoning chains, tool invocations and outputs — sufficient to reconstruct why an action was taken, not merely that it was.

●        Lifecycle inventory. From proposal through decommissioning, linked to the business capability the agent serves.

On the survey evidence, two figures are worth separating because they are routinely merged. A SailPoint vendor survey of technology professionals reported that around 80% of companies said their AI agents had taken unintended actions, including accessing unauthorised systems and sharing sensitive data.¹³ A separate Deloitte survey of 3,235 director-to-C-suite respondents across 24 countries found only about 21% reporting a mature agent governance model.¹¹ These come from different populations with different methods, and the SailPoint figure comes from a vendor selling identity governance — an interest directly served by the finding. Both are plausible and neither is incident data; they measure what respondents report perceiving. The governance gap is real, but the precision of these numbers should not be over-read.

Regulatory constraints are firmer ground. The EU AI Act and the NIST AI Risk Management Framework impose risk-tiered inventories, approval workflows and post-market monitoring obligations. Data sovereignty requirements increasingly determine where workloads and models may run. These are architectural constraints with legal force, which makes them among the few parts of this agenda that will not be negotiated away under delivery pressure.

What this asks of architects — and of managers

The demands run in both directions, and the second direction is usually understated.

Architects need to shift from gatekeeping to enablement, which is easy to say and involves a genuine loss of formal authority. It also means adopting the vocabulary of management — financial performance, risk appetite, customer outcomes — rather than expecting executives to learn service meshes. And it means a real competency in behavioural and organisational analysis, not as an adjunct but as a core method: if the social system determines whether the architecture takes, then reading the social system is architectural work.

Managers and executives need corresponding technical fluency, and this is where most organisations are weakest. Three capabilities in particular: understanding business functions as composable, API-addressable services rather than as departments; treating data semantics as an operational asset that requires executive decisions about definitions and ownership; and being able to set risk thresholds, autonomy boundaries and auditability requirements for automated decisions. That last one cannot be delegated to the architecture function. Deciding how much autonomy a system may exercise over a customer decision is a business judgment with legal consequences.

The asymmetry matters. EA has spent three decades asking architects to learn the business. The agentic shift makes the reverse requirement urgent, and no certification track currently addresses it well.

Where this leaves us

Enterprise architecture’s paradox is real: a discipline built by engineers, applying engineering models to organisations that are not machines. But it is not novel, and framing it as EA’s peculiar affliction obscures the more useful reading. Trist and Bamforth established the underlying result in 1951, Leavitt formalised the components in 1965, Conway inverted the causal arrow in 1968. Architecture practice has had the answer available for most of its existence.

The agentic turn does not settle the question. It raises the cost of getting it wrong. When execution runs through people, a flawed model is buffered by human discretion; when it runs through agents, the flaw executes. That makes the architectural task more consequential and, if anything, harder — because it now requires getting the semantics right in advance, at machine precision, for processes whose ambiguity was previously doing useful work.

The practical implication is unglamorous and fairly clear. Design the social system and the technical system together, or accept that you have designed neither. Model to answer questions rather than to achieve coverage. Govern agents as identities with scoped authority rather than as features. Argue in the language of the people who fund the work. And treat the informal organisation not as noise obscuring the architecture, but as a substantial part of what the architecture actually is.

References

1. Trist, E. L., & Bamforth, K. W. (1951). Some Social and Psychological Consequences of the Longwall Method of Coal-Getting. Human Relations, 4(1), 3–38. The founding study of socio-technical systems theory.

2. Leavitt, H. J. (1965). Applied Organizational Change in Industry: Structural, Technological and Humanistic Approaches. In J. G. March (ed.), Handbook of Organizations. Source of the task–structure–people–technology model.

3. Conway, M. E. (1968). How Do Committees Invent? Datamation, 14(5), 28–31. http://www.melconway.com/Home/Committees_Paper.html

4. Zachman, J. A. (1987). A Framework for Information Systems Architecture. IBM Systems Journal, 26(3), 276–292.

5. Forrester, Enterprise Architecture Is Still Not Getting The Recognition It Deserves (February 2023), drawing on the Modern Technology Operations Survey 2022. https://www.forrester.com/report/enterprise-architecture-is-still-not-getting-the-recognition-it-deserves/RES178892

6. Avolution, Enterprise Architecture in 2026: Trends, Skills, and Strategies. Source of the widely circulated 32% figure; cited here as the origin of the claim, not as evidence for it. https://www.avolutionsoftware.com/our-resources/enterprise-architecture-trends-skills-strategies-2026/

7. Quinn, R. E., & Rohrbaugh, J. (1983). A Spatial Model of Effectiveness Criteria. Management Science, 29(3), 363–377. See also Cameron, K. S., & Quinn, R. E. (1999/2011), Diagnosing and Changing Organizational Culture.

8. Edmondson, A. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly, 44(2), 350–383.

9. Hofstede, G. (1980/2001). Culture’s Consequences. On power distance as a cultural dimension.

10. McKinsey Global Institute (June 2023). The Economic Potential of Generative AI: The Next Productivity Frontier. Source of the $2.6–$4.4 trillion estimate across 63 use cases. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier

11. Deloitte (2026), State of Generative AI in the Enterprise survey wave; N = 3,235, director to C-suite, 24 countries. Source of the 21% mature agent governance and 84% roles-not-redesigned figures.

12. Bainbridge, L. (1983). Ironies of Automation. Automatica, 19(6), 775–779.

13. SailPoint (May 2025), AI agent adoption and security research. Vendor survey of technology professionals; source of the 80% unintended-action figure. https://markets.financialcontent.com/wral/article/bizwire-2025-5-28-sailpoint-research-highlights-rapid-ai-agent-adoption-driving-urgent-need-for-evolved-security

14. A Socio-Technical Analysis of Enterprise Architecture Misalignments. International Journal of Advanced Computer Science and Applications. https://thesai.org/Downloads/Volume17No3/Paper_87-A_Socio_Technical_Analysis_of_Enterprise_Architecture.pdf

15. Enterprise architecture: beyond business and IT alignment. Innovation & Regulation Chair, Telecom Paris. https://test-innovation-regulation.telecom-paris.fr/wp-content/uploads/2017/12/DEDM13_Enterprise-architecture-beyond-business-and-IT-alignment.pdf

16. Bückle, S. et al. Socio-technic Dependency and Rationale Models for the Enterprise Architecture Management Function. Technical University of Munich (sebis). https://www.cs.cit.tum.de/fileadmin/w00cfj/sebis/publications/2011/Bu11e.pdf

17. McKinsey, Rethinking Enterprise Architecture for the Agentic Era. https://www.mckinsey.com/capabilities/mckinsey-technology/our-insights/rethinking-enterprise-architecture-for-the-agentic-era

18. Architecture Decision Record (ADR) organisation and templates. https://github.com/architecture-decision-record/architecture-decision-record

19. INNOQ, Socio-Technical Architecture as a Competitive Advantage (2025). https://www.innoq.com/en/blog/2025/04/soziotechnische-architektur-als-wettbewerbsvorteil/

20. European Union Artificial Intelligence Act (Regulation 2024/1689); NIST AI Risk Management Framework (AI RMF 1.0, January 2023).