Pricing Consulting Services in the Age of Agentic AI
digitaleconomy.stanford.edu · 24 October 2026
Brynjolfsson and Petropoulos (August 2026) document that consulting firms historically priced work using time-and-materials or project-based fixed fees, with pricing models evolved to manage information asymmetries between clients and consultants. The paper states that AI agents now automate analysis, generate insights, standardise workflows, and interact directly with clients, blurring boundaries between advisory service, software product, and managed operation. Consulting firms now sell access to internal AI platforms, productized AI agents embedded in client environments, and hybrid arrangements.
This is the observable fact the cited source supports, stated without interpretation.
Testing LAW_2 (value moves from access toward outcomes). The explicit documentation that consulting firms are moving from time-based and project-based pricing toward hybrid and outcome-based structures, driven by AI agent deployment, directly addresses whether the mechanism is observable in a commercially significant sector. [Strength capped from 4 to 3: source tier 3.]
This directly addresses LAW_2 (value moves from access toward outcomes) and the restructuring of pricing models away from pure access-based structures. The paper documents that the classical price equation (number of consultants times billable hours times rate per hour) no longer describes what consulting firms actually do and charge for. This is consistent with a transition from access-based pricing (paying for consultant time availability) to hybrid and outcome-based structures. The embedding of AI agents in client operations and the sale of internal AI platforms represent a shift in the locus of value capture from the consultant's time to the productivity and outcomes of the AI system.
This directly addresses LAW_2 (value moves from access toward outcomes) and the restructuring of pricing models away from pure access-based structures. The paper documents that the classical price equation (number of consultants times billable hours times rate per hour) no longer describes what consulting firms actually do and charge for. This is consistent with a transition from access-based pricing (paying for consultant time availability) to hybrid and outcome-based structures. The embedding of AI agents in client operations and the sale of internal AI platforms represent a shift in the locus of value capture from the consultant's time to the productivity and outcomes of the AI system.
This is the framework's reading of the fact above. The source does not endorse it and is not cited as doing so.
This academic paper documents the state of consulting industry pricing practices as of August 2026. It observes that firms are already in transition away from pure time-based pricing. The paper's conceptual framework suggests that AI agents are pushing consulting firms toward hybrid and asset-based structures. The test for LAW_2 is whether value moves from access toward outcomes over 1 to 2 years post-credible deployment. The consulting industry has been subject to AI agent deployment since approximately 2022 to 2023, placing this observation within the relevant horizon. However, the text does not provide quantitative evidence of durability or margin pressure on the old models.
Current evidence is recorded immediately. The paper's longer empirical horizon determines when persistence becomes a stronger framework-level test.
The paper is a framework-building exercise drawing on examples from leading firms and confidential expert discussions. It does not provide empirical evidence of widespread pricing model transition or quantitative data on the success or failure of different models. The documentation of AI capability (agents automating analysis, standardising workflows) does not establish that these capabilities have fundamentally changed the economic structure of the consulting business, only that they have created new options for service delivery.
Recorded for every observation assessed at strength 4 or 5. A record that cannot state the strongest competing reading of its own evidence is not published.
Where several outlets report the same development, they are recorded as additional sources on this one observation and are not counted as separate evidence.