A live empirical record. Every observation below is a market development that tests the prediction, drawn from the Autonomy Economics Evidence Monitor and reviewed before publication.
Four indicators
Commercial-unit pressure
5observations
Are buyers translating lower human effort into pressure on the unit a firm charges for?
3 supportive · strongest observation
Pricing redesign
8observations
Are firms moving value capture from access toward outcomes?
7 supportive · strongest observation
Interface displacement
10observations
Are agents replacing the human as the interface to economic activity?
4 supportive · strongest observation
Revenue durability
10observations
Is the durability of existing commercial structures being repriced by customers or capital?
5 supportive · strongest observation
Counts are observations recorded against each indicator, not a score. A count rises because something happened in a market, not because a view hardened.
The mechanism
The paradox is not that AI fails. It is that AI works, and the commercial structure it works inside was built for a different kind of activity.
Consulting
AI capability rises
↓ Human progression falls
↓ Client insourcing rises
↓ Billable work falls
↓ Pricing pressure rises
Software
Agent activity rises
↓ Human interface dependence falls
↓ The seat becomes a weaker proxy for value
↓ Usage and outcome pricing rise
Legal
AI productivity rises
↓ Required lawyer hours fall
↓ The client recognises the saving
↓ The hourly fee is challenged
Redesign, or not
The prediction is not that every exposed firm declines. It is that firms which change how they capture value alongside the change in activity can hold durability, and firms which leave the commercial unit tied to the old activity structure cannot. Both sides are on the register.
16 observations test the paradox directly. Each states the fact its source supports, the prediction it tests, and the strongest competing reading of the same evidence.
In a September 2026 investor presentation, Blackstone President and COO Jon Gray described substantial realised AI productivity gains across portfolio companies, then highlighted multiple compression in professional services, software and information-services businesses even where underlying operations remained sound, and said management teams that move the model from seats to outcomes should be better positioned.
What it tests
Whether autonomous capability can generate substantial operating gains while markets simultaneously question the durability of legacy commercial structures, and whether the proposed response shifts value capture away from seat-based access toward outcomes.
The strongest argument against this reading
The productivity examples and the repriced public companies are not necessarily the same businesses, so the presentation does not establish a firm-level causal chain from AI productivity to valuation compression. Multiple compression and lower software deal activity can also reflect interest rates, macro conditions or ordinary technology-cycle uncertainty. The seat-to-outcome statement is a forward-looking management view and does not show that such pricing has already been adopted at scale.
Salesforce documentation states that Help Agent pricing is strictly outcome-based. Unresolved sessions are not billed, while successful resolutions are billable outcomes.
What it tests
As autonomous systems perform work, value capture can migrate from access and activity units toward successful outcomes.
The strongest argument against this reading
Outcome pricing may remain limited to narrowly measurable customer-service use cases.
Atlassian states that Customer Service Management AI agent resolutions will be charged at $1 per successful resolution from 3 December 2026, only when the AI fully resolves a customer request without handing it to a human.
What it tests
Autonomous software can move value capture from user access toward successful outcomes, with verification determining when the economic event has occurred.
The strongest argument against this reading
The model is announced but billing has not begun, and customer service is unusually easy to measure.
The FT reported continued high redemption requests at Blackstone’s large retail private-credit fund and linked investor concern partly to exposure to leveraged software companies facing uncertainty over AI’s impact, while also identifying broader private-credit pressures.
What it tests
Autonomy exposure can influence valuations and capital structures before operating revenue fully resets, including through credit and portfolio channels.
The strongest argument against this reading
Redemptions have multiple causes, including liquidity preferences, fund structure, leverage and general private-credit sentiment; AI cannot be isolated as the cause.
Demand for law firm hours rose 4.2% through the first half of 2026 against a normal rate of about 1.5%, while two-thirds of firms reported daily associate use of AI and clients pressed to move away from the billable hour.
What it tests
Tests whether realised changes in human progression are visible alongside pressure on effort-based pricing, while recognising that aggregate hours are not the same measure as effort per unit of work.
The strongest argument against this reading
Hours could be rising because deal volume is rising, an M&A revival raises demand independently of how long each matter takes, so total hours may grow even while hours per unit of work fall. The measure reported is aggregate demand, which differs from effort per matter, and the framework's claim is about effort per matter. On that reading the record says little about progression and the mechanism remains untested here.
The FT reported that Morgan Stanley, Citigroup and Goldman Sachs are pressing major law firms to reflect AI-enabled efficiencies in fees, including through competitive bidding, fixed fees and requests for evidence of AI-driven savings.
What it tests
Commercial-unit pressure should appear once buyers can observe that less human progression is required, even if demand for high-value legal judgement remains.
The strongest argument against this reading
Large financial institutions may have exceptional bargaining power and may not represent the broader legal market.
The FT reported that PE-backed software companies are using shorter amend-and-extend transactions, paying higher yields and accepting stronger creditor protections as lenders assess AI-related threats to long-term software business-model durability alongside leverage and maturity risks.
What it tests
Markets and creditors can reprice the durability of exposed commercial structures before full operating-model or revenue deterioration is visible.
The strongest argument against this reading
Leverage, maturity walls, interest rates and CLO dynamics may explain much of the financing pressure independently of AI.
The FT reported that companies are bringing more technology work in-house, reducing reliance on large consulting teams and demanding lower, fixed or performance-linked fees as AI reduces the human effort required for implementation and analysis.
What it tests
Consulting can remain in demand while billable progression becomes less defensible and value moves toward senior judgement, leadership advice and measurable outcomes.
The strongest argument against this reading
Consulting pressure may also reflect procurement cycles, macro conditions and broader cost-cutting.
Atlassian documents Rovo credits as a meter for AI usage and enriched Teamwork Graph context, including calls made through the Rovo MCP server and CLI.
What it tests
As agents, in place of users, consume software capability, the seat becomes a weaker standalone measure of value and new activity and context units emerge.
The strongest argument against this reading
Usage pricing may simply supplement seats, leaving them neither replaced nor weakened.
Vercel's September 2026 AI Gateway Production Index shows open-weight models processing 56% of gateway token volume in August while accounting for 14% of estimated spend. Average price per token fell 23.2% during the month, while Anthropic retained 64% of gateway spend. The same dataset therefore shows both rapid diffusion of lower-cost capability and persistent concentration of premium spending.
What it tests
Tests whether broad availability of cheaper and open-weight capability erodes scarcity rents, or whether premium frontier capability continues to sustain concentrated spending.
The strongest argument against this reading
Vercel AI Gateway is one routing platform and does not represent the whole enterprise AI market. Estimated spend uses public list prices in place of actual negotiated bills, and differences in workload complexity can justify premium model usage without establishing durable scarcity rents.
Salesforce and Anthropic announced Claudeforce, making Salesforce data, workflows, business logic, actions and governance accessible directly in Claude, initially through a plugin with 37 prebuilt sales skills.
What it tests
External agents can become the user-facing interaction layer while the incumbent protects economic relevance by becoming the governed execution and context layer underneath.
The strongest argument against this reading
The partnership may increase Salesforce engagement without materially changing the role of seats or interfaces.
Salesforce reported $10.8 billion in subscription and support revenue, up 12% year on year, nearly $3.9 billion in Agentforce and Data 360 ARR, 7 billion Agentic Work Units delivered to date, and reorganised revenue disclosure around Agentforce Apps and Data 360, Headless Platform and Other.
What it tests
Firms that realign commercial structure with autonomous activity can remain economically durable while interface dependence and conventional activity units change.
The strongest argument against this reading
Growth may be explained by acquisition effects, installed-base strength and conventional subscription expansion.
Salesforce expanded Headless 360 across its platform, exposing reusable enterprise capabilities through MCP, developer tools and headless experiences so authorised AI agents can discover and take action without relying on conventional application interfaces.
What it tests
As agents become an interaction layer, interface dependence weakens and durable value can migrate toward underlying systems of record, workflow and governed action.
The strongest argument against this reading
Headless access may complement conventional user interfaces instead of displacing them.
Deloitte’s survey of 121 senior legal leaders found 61% in AI deployment phases, 61% experimenting with or piloting agentic AI, 78% wanting external-provider AI to reduce costs, and 85% expecting AI to change law-firm pricing.
What it tests
When structured legal production requires less human progression, billable-hour durability should come under pressure while judgement and accountability remain more defensible.
The strongest argument against this reading
Survey expectations may not translate into realised fee changes.
Salesforce reported record first-quarter results, more than $1 billion in Agentforce ARR, $3.4 billion in combined AI and data ARR, 3.8 billion Agentic Work Units delivered, and more than one million active Slack MCP users within six weeks of launch.
What it tests
Firms that redesign around autonomous activity can preserve durability even while the old interaction structure weakens.
The strongest argument against this reading
Salesforce may simply be benefiting from scale, installed-base strength and bundling, with agentic redesign playing little part.
Thomson Reuters' 2026 Future of Professionals survey, conducted in March and April before the Autonomy Economics publication date, found a clear gap between client expectations and law-firm pricing response. Seventy-one per cent of in-house legal professionals expected outside firms to change their commercial models as AI usage increased, while only 28% of law firms reported pricing changes and 62% said pricing remained unchanged. Because the underlying survey predates publication, this record is treated as contemporaneous background and does not count as post-publication evidence.
What it tests
Tests whether AI-enabled changes in professional work translate into changes in effort-based commercial units and whether commercial structures adapt alongside activity structures.
The strongest argument against this reading
The survey measures expectations and reported pricing changes. It does not measure realised reductions in human progression or long-run revenue durability. Law firms may be adapting through rates, staffing, scope or service mix without changing the formal pricing structure.
The Agentic Profit Paradox is part of the Autonomy Economics research framework. Evidence is maintained through the Autonomy Economics Evidence Monitor, which records the full register across every mechanism, law and sector, including the evidence that runs against the framework.
A further 8 observations are relevant background rather than direct tests of the paradox: capital-coordinated deployment structures and the governance layer around autonomous systems. They are recorded in full on the monitor.
Nothing here is published automatically. Every record was reviewed by a person, who checked the source, the separation of observed fact from interpretation, and the chronology. Read the methodology or the evidence that could challenge the framework.