The Role of Designers in the AI Era — Industry x Academia Roundtable (Business-Only Revision)

Revision history: The first version had industry personas citing academic literature and using Evidence Ledgers, making it skew academic. The critique was that “people in industry do not have knowledge of academic papers; they would argue in terms of whether it is profitable, whether it can be explained to and billed to clients, and whether AI can substitute their own work for efficiency.” Industry personas were redesigned as business-only (only three lenses: money, billing, and efficiency; no scholar names, theories, or academic terminology), and the roundtable was re-conducted. Theoretical connection is handled exclusively by academic critics (R4). This page is the business-only revision that replaces the previous version.

Industry personas (practice, cost-benefit): F = design-firm-lead (IDEO/frog type) / C = strategy-consulting-design-lead (McKinsey/Song type) / O = independent-design-office-lead (Pentagram/boutique type). Academic critics (theory): Theory / Education / Industry. Business evidence ledger: industry.md / voices.md. The primary deliverable is a divergence table with citations, not a converged conclusion. Related: ai-in-design-industry / ai-in-design-literature / ai-in-design-2026 / the roundtable on delivering multi-agent value / ai-design-near-term-flashpoints (near-term discussion hub) / the roundtable on design consulting and generative AI (sequel: focus on coexistence strategies).

Question and Method (Role Assignment)

Question: “What role will designers play in the AI era?” Industry personas argue exclusively from cost-benefit (is it profitable? / can it be billed to clients? / can AI replace their work for efficiency?). Academic critics connect those business claims to theory and adjudicate. This establishes a clear division of labor: “industry = on-the-ground financial calculus” and “academia = theoretical reasoning.”

TL;DR

Even when three industry parties argue exclusively from financial calculus, the conclusions point in the same direction as the previous version, but more sharply: “roles will persist / move up” is a sales pitch without unit-price evidence, and all three agree that handing AI the entry-level work (production) will starve the talent pipeline. Academic critics translate this purely business observation into theory, showing that industry’s cost-benefit calculus independently rediscovers the core of academic findings while simultaneously exposing their weaknesses.

Key Points from R1 (Industry, Business-Only)

• F: Billable production hours will certainly decrease -> this usually means unit price decline, not “role elevation” (only firms that converted to outcome-based billing maintain margins). “Selling taste/edges” is directionally correct but no unit-price data exists for proposals where taste was accepted as a line item = billing basis absent. Selling capability (“teach them to fish”) kills next-year repeat business. Biggest management risk = junior work disappears -> talent pipeline for taste dries up.
• C: ROI (+32pp) is ammunition for proposals but is correlation, not causation; passing it off as causal kills repeat business at year 3. The 16% wall (only 16% fully redesigned) means PoCs evaporate before enterprise-wide rollout. Product builder under human-hour billing with smaller teams = own revenue decline + junior layoffs (without converting to value-based billing, you cut your own unit price). Efficiency gains ≠ revenue growth (the utilization trap). ROI-only positioning misses differentiated engagements (A: ROI-proposal type vs. B: craft-premium type; we can only sell A).
• O: Authorship commands premium pricing only when the engagement is a named personal commission; studio-name authorship thins under AI and loses to cheaper competitors. Craft is also compromised by client budgets (Scher: “there was no budget”). AI replaces handwork, improving margins, but handing over the core (initial style/sketch) invites “we can do this ourselves” and destroys the named commission -> unit price decline. Without billing for IP and liability, small teams face bankruptcy. Scher (budget-pragmatist, AI-adopter) vs. Sagmeister (selling authenticity) are the two poles.

Items to Confirm (Practical Data Needed Next)

• Unit prices for engagements where “taste fee” was accepted as a proposal line item / Production labor reduction rate, utilization rate, and proposal win rate after AI adoption (F, uncompiled both internally and industry-wide).
• PoC-to-enterprise rollout yield (actual numbers behind the 16% wall) / Post-AI-adoption gross margin / 3-year repeat rate for ROI-driven engagements (C).
• Conversion data from non-prominent designer authorship to income / Relationship between AI usage rate and unit pricing (O).
• Reinvestment rate of efficiency gains and its relationship to talent reproduction (to resolve the distribution funding divergence).
• Corpus items to confirm: C-01 (McKinsey “Designing for Machines” primary source not reached) / SC-04 (excluded) / NC State “50%” original source / Exeter n=500 and some percentages (via secondary source). DF-01 is a single full-service firm.

References

Industry personas argued business-only, citing zero scholars, theories, or papers (business evidence = figures, engagements, and client reactions are recorded with URLs in industry.md / voices.md). The canons below were all used by academic critics (R4) for theoretical connection, and many are secondary references from memory (pages unconfirmed = [to be confirmed]). Primary verification is needed before citation. Protocol: .claude/practice-roundtable-protocol.md (practice-first) / .claude/critique-protocol.md (academic critics only).

• Abbott, A. (1988). The System of Professions. --- jurisdiction / subordinate jurisdiction (de-economization of jurisdiction is an extended concept from this roundtable, not Abbott’s term). [to be confirmed]
• Braverman, H. (1974). Labor and Monopoly Capital. / Friedman (1977). Industry and Labour (responsible autonomy) / Burawoy (1979). Manufacturing Consent. --- deskilling and its counterarguments. [to be confirmed]
• Bourdieu, P. (1993). The Field of Cultural Production / (1979). Distinction. --- symbolic capital, capital conversion. Market closure follows the Weber/Parkin lineage (distinct from Bourdieu). [to be confirmed]
• Kimbell, L. (2011/2012). “Rethinking Design Thinking” I/II. Design and Culture. --- critique of management buzzwordification. [to be confirmed]
• Ries, E. (2011). The Lean Startup. --- validated learning (learning velocity over utilization rate / tension with utilization-rate legitimation). [to be confirmed]
• Lave, J. & Wenger, E. (1991). Situated Learning. --- legitimate peripheral participation (trajectory origin). [to be confirmed]
• Dreyfus, H. & Dreyfus, S. (1986). Mind over Machine. --- five-stage skill acquisition model (upper stages are sedimentation of lower-stage experience). [to be confirmed] (status: contested)
• Vygotsky, L. (1978). Mind in Society. --- internalization, mediation. [to be confirmed] / Papert, S. (1980). Mindstorms. --- constructionism / objects-to-think-with, ownership. [to be confirmed]
• Kirschner, P., Sweller, J. & Clark, R. (2006). “Why Minimal Guidance During Instruction Does Not Work.” Educational Psychologist 41(2). --- worked example (fading required). [to be confirmed]
• Schon, D. (1983). The Reflective Practitioner. / Polanyi, M. (1966). The Tacit Dimension. / Cross, N. (1982/2006). “Designerly Ways of Knowing.” / Simon, H. (1969). The Sciences of the Artificial. --- the existence of design knowledge that resists translation / modes of design knowledge. [to be confirmed]
• Buchanan (1992) & Rittel & Webber (1973) wicked problems / Krippendorff (2006) The Semantic Turn (agency in meaning-making) / Costanza-Chock (2020) Design Justice (who designs). [to be confirmed]

Update Policy

This note is a living page. When practical data from the confirmation queue is obtained, it will be appended and updated will be revised. Further re-discussions will be added as subsequent rounds. Industry persona apparatus: .claude/skills/designer-role-roundtable (convening) + each perspective skill (business-only = .claude/practice-roundtable-protocol.md).