Generative and agentic AI are reanimating the ambition of 1990s business‑process reengineering under new, legitimate vocabularies — unlocking bigger redesign opportunities but reintroducing familiar failure modes and distributional risks.

Main Finding

AI’s recent generative and agentic capabilities are creating conditions that make elements of 1990s business process reengineering (BPR) logically and practically attractive again — but not as a simple revival of the old label. Instead, BPR’s core ambition (radical process redesign enabled by new technology) is being “translated” into adjacent, more legitimate vocabularies (AI‑augmented BPM, Large Process Models, agentic BPM). This translated resurgence changes technical feasibility but leaves many classic BPR risks and uncertainties intact.

Key Points

• Historical context: BPR rose as a rhetorically bold prescription for radical redesign in the 1990s, later fell from legitimacy after many implementations failed or produced disruptive side effects (downsizing, managerial overreach).
• The paper is conceptual: it revisits BPR through the lens of management fashion theory rather than presenting new empirical tests.
• Management-fashion lens: organizational ideas cycle; legitimacy, rhetoric, and vendor/consultant ecosystems shape adoption. “Translated resurgence” is introduced to capture reappearance of older ideas under new labels and reinterpretations.
• Translation targets today: business process management (BPM), AI‑augmented BPM systems, Large Process Models (LPMs) that map processes and knowledge, and agentic BPM where autonomous agents drive process steps.
• AI’s contribution: generative and agentic AI can (a) automate cognitive and coordination tasks; (b) infer, model, and recompose process knowledge from data and text; (c) orchestrate multi‑step workflows via agents — thereby materially expanding what is technically possible for redesign.
• Persistent risks: technical feasibility improvements do not eliminate classic problems — implementation failure, hidden organizational interdependencies, workforce displacement, managerial overreach, legitimacy problems — and may amplify some harms.
• Outcome: the situation is best seen as an emerging, unsettled phase in BPR’s afterlife: selective reactivation of its logic through new vocabularies and technologies, producing both genuine redesign potential and renewed implementation hazards.

Data & Methods

• Type of paper: perspective/conceptual article (no systematic literature review or new empirical analysis).
• Primary theoretical lens: management fashion theory (how ideas gain/lose legitimacy and are reinterpreted).
• Method: conceptual synthesis of BPR history, process‑management scholarship, and recent AI capabilities and industry discourse. Introduces the analytical concept “translated resurgence.”
• Evidentiary basis: discourse and research‑literature mapping rather than primary datasets or formal empirical tests.

Implications for AI Economics

• Firm behavior and productivity
  • Potential for larger, non‑marginal redesigns that change task boundaries and work organization — increasing potential productivity gains but also increasing implementation risk and heterogeneity in realized returns across firms.
  • Productivity estimates from early adopters may be biased upward if selection and managerial fashion effects are ignored.
• Labor markets and tasks
  • Greater automation of cognitive/coordination tasks could shift labor demand across occupations and skill groups; displacement risks remain substantial where redesign combines automation with headcount reductions.
  • Complementarities: gains may accrue to firms/workers who successfully redesign roles around AI (skill upgrading, supervision of agents), widening within‑ and between‑firm inequality.
• Investment, rents, and market structure
  • Vendor ecosystems (consultants, BPM/AI platform providers) and managerial signaling can magnify diffusion even where expected returns are uncertain; rent capture by platforms/consultants is a plausible outcome.
  • Large incumbents with data/process scale may secure advantages from agentic BPM and LPMs, potentially increasing concentration.
• Measurement and empirical strategy
  • Need for process‑level data: adoption of AI‑BPM tools, audit trails from BPM systems, time‑use and task data, and firm‑level reengineering investments to credibly estimate causal effects.
  • Recommended methods: randomized pilots or staged rollouts, difference‑in‑differences with careful controls for managerial fashion timing, structural models accounting for adoption costs and failure probabilities, and mixed methods (detailed case studies + microdata).
• Policy and welfare
  • Policy should account for transitional costs (retraining, unemployment spells) and potential market power effects from platformized process models.
  • Regulation and governance needed for agentic systems that can autonomously change workflows (auditability, safety, accountability, labor protections).
• Research agenda (concise bullets)
  • Quantify the frequency and scale of “radical” vs incremental redesigns in AI‑enabled projects.
  • Measure heterogeneity in returns and identify organizational factors (capability, governance, vendor reliance) that predict success vs failure.
  • Model the general equilibrium effects of widespread process redesign (task reallocation, wage composition, market concentration).
  • Evaluate policy interventions: retraining, redistribution, and rules for agentic process changes.

Summary conclusion: AI has plausibly reactivated the core logic behind BPR via new technical capabilities and legitimizing vocabularies, but this translated resurgence creates both renewed opportunities for large productivity gains and familiar — sometimes amplified — implementation and distributional risks. Empirical work that focuses on process‑level change, organizational capabilities, and adoption dynamics is needed to assess net economic effects.