AI Commoditization Resilience of Design Billing Models — Output-Based Billing Is Most Vulnerable, and Retainer Stickiness Does Not Mean High Margins — Notes

AI resilience of billing models ranks 'outcome-based > time/human-hour > output (deliverable)' (F and C diverge on relative vulnerability of human-hour vs. output billing).

Claim (one sentence): In a market where AI can produce equivalent deliverables in hours, “the moment you define billing in terms of a thing, the price slides to AI’s marginal cost.” Only what cannot be objectified as a thing (decision-making, accountability, operations, named referrals) can be defended, but the difficulty of replacing a retainer means ‘hard to fire,’ not ‘billable at a premium.’

Origin: AI Flashpoints Certain to Arrive in the Next 1-3 Years — Three Design Leads' Near-Term Flashpoints (2026-2029) R3 (danger of output-based billing), R4 (prescriptions), R5 (cost-benefit assessment of SO prescriptions). Relationship to democratization: The Democratization-Recommodification Paradox: How AI Turns 'Anyone Can Do It' into 'No One Can Charge for It'. Industry data: AI Adoption in the Design Industry: Perspectives from Policy, Data, and Product Developers (2026). The legal structure whereby copyright gaps and disclosure obligations erode billing rationale: EU AI Act Article 50 and Design Practice — A Structural Analysis on the Eve of Enforcement.

AI Resilience of Billing Models (Three-Party Assessment)

	Output (Deliverable) Billing	Time/Human-Hour Billing	Outcome-Based Billing
F (Firm)	Weakest (thing = reproducible = on the same shelf as AI; floor drops faster than human-hour)	Medium	Strongest (but viable engagements are scarce)
C (Consulting)	Medium-weak	Weakest (labor reduction = immediate price cut)	Strongest (but pure outcome-based is rare; dependent on fixed fees)
O (Independent)	Defenseless without a named referral / a different game with one	---	---

The crux: the moment a deliverable is defined as a thing, the price slides to AI’s marginal cost (near zero). F and C diverge only on which is more dangerous between human-hour and output billing (non-convergence).
“Output-based billing” splits into work-based billing (who made it = defensible via named referral) and deliverable billing (what was delivered = defenseless) (O’s self-correction). The majority of contracted work falls into the latter.

The Retainer Trap --- Difficulty of Replacement ≠ High Margins

Do not conflate switching cost (hard to fire) with gross margin (billing at a premium) (three-party consensus).
F: “Stickiness is a shield that buys time against price cuts. Retainer gross margins fall to 20-35% [to be confirmed], and the game becomes low churn and LTV. Profitability requires offloading operations to cheaper headcount + AI = headcount reduction as the funding source for breaking even.”
C: “Seat-based billing is thin-margin human-hour work; AI efficiency gains flow to the client and do not scale. The exit (client completes in-housing, cuts out the consultant) arrives before profitability.”
O: “Profitability is an all-in bet on whether vertical specialization with named referrals materializes. If it does not, increasing the number of engagements turns into hourly labor that runs red.”

The “Gatekeeper Business” Will Collapse --- Unit Pricing Splits by an Order of Magnitude

Upper tier: AI governance framework design (policies, responsibility demarcation, audit) = transformation engagements worth hundreds of millions = consulting captures this (not the business at this layer).
Lower tier: Visual inspection of AI output = quasi-delegation inspection work = a self-shrinking business that contracts as AI accuracy improves = floor pricing.
O: “Clients pay for ‘who guarantees this’ (same as accountants and lawyers), but the gatekeeper bifurcates into named accountability guarantors (sellable at a premium) and anonymous inspectors (floor pricing).” “Redefining yourself as a gatekeeper will keep you afloat” conflates those who capture the upper tier and those who fall to the lower tier under the same label.

Unresolved

(Volume x surviving margin) = the numbers that determine survival at this layer, and none of the three parties have actual data (AI Flashpoints Certain to Arrive in the Next 1-3 Years — Three Design Leads' Near-Term Flashpoints (2026-2029) R4).
Post-retainer-transition churn rate, gross margin rate, and fixed-cost coverage ratio (F’s 20-35% is gut feeling [to be confirmed]).
Three-year trajectory of gatekeeper (accountability guarantee) unit pricing and its commoditization velocity.

References

Origin: AI Flashpoints Certain to Arrive in the Next 1-3 Years — Three Design Leads' Near-Term Flashpoints (2026-2029) R3-R5 (cost-benefit assessment by design leads F/C/O. This note extracts and reorganizes that assessment)
External sources (digitalapplied / Simon-Kucher / Net-Craft etc.) are recorded with URLs in the references section of the parent note AI Flashpoints Certain to Arrive in the Next 1-3 Years — Three Design Leads' Near-Term Flashpoints (2026-2029). All are agency-side publications and are discounted as position talk (speculative URL entries are avoided)