AAI snapshot · Overall readiness
2026-W46 · scale 0–100
Awareness
0–20Active
21–40Operational
41–60Systemic
61–80Transformational
81–100
Organization
72
Brand
44
Product
68
Composite = unweighted average of its five dimensions. Overall = unweighted average of the three composites. Bands computed at render.
Evidence this issue · 24 cited sources across 30 signals · 10 primary-source (42%) · 15/15 dimensions carry a cited signal · 6 dimensions moved
¶
This week
What moved the register
This week the index tells one story from two directions, and both are true at once. At the product layer, the building blocks got better. Google shipped a Gemini Decisions API for calibrated multimodal decisions, OpenAI put a governed payment rail behind agents with user approval in the loop, and Microsoft Research demonstrated formal verification of agent-written code against developer intent. The interaction layer, the agent-addressable infrastructure, and the trust tooling all advanced. Three dimensions moved up on cited evidence.
At the organization layer, the week was a correction. The HFS and Rackspace survey of 502 senior decision-makers is the hinge document: only 33 percent of enterprises met or beat their biggest AI program's business case, and that drops to 28 percent once AI spans several functions. The same survey puts 96 percent reporting significant governance gaps, with 78 percent calling governance failure a bigger threat to scaling than technology failure. Kyndryl's read that 64 percent of leaders report positive ROI while readiness fell says the same thing from the inside: satisfaction is running ahead of the operating model that would justify it.
The sharpest signal is agentic. Sinch found 74 percent of organizations with live customer communications agents had rolled back or shut down a deployment after governance failures, and Gartner expects over 40 percent of agentic projects canceled by end of 2027. This is not a messaging problem. It is the gap between a demo that works and an enterprise that can own the work after automation, at a cost and risk it can live with. The Superblocks field study names the culprit plainly: legacy systems and absent data pipelines block agents that look finished in the demo.
Read through the Compiled Corporation lens, the pattern is clean. Capability is compounding at the product layer. The firm's ability to put that capability into its own decision-making is where the reversals happen. The fault line is governed data and named ownership: only 16.9 percent of governance measures have an explicit owner, and 44 percent of enterprises have no consistent way to run AI in production. An agent that can act on a payment rail is worthless to a firm that cannot say who owns the workflow when it acts wrong.
The regulatory floor rose this week too. New York's RAISE Act opened its registration portal with penalties up to $3M per subsequent violation. The Identity Control Surface is no longer optional architecture. It is becoming the registration requirement.
The net: three product dimensions up, three organization dimensions down. A split week is an honest week. The firms winning the next quarter are the ones treating that 74 percent rollback figure as a specification, building ownership, evidence, and runtime controls before the agent touches production. The capability is already here. The architecture to hold it is the work.
¶
Score movement · week over week
What moved this week
6 dimensions shifted. The trend is the product.
▲ Brand · Agent-Ready Infrastructure
Agent-Ready Infrastructure
+1 · 62 → 63
ChatGPT Wallet puts a governed payment rail behind agent purchases with user approval and card protections, and Menlo documents agent infrastructure scaling as a distinct spend category. The agent-addressable layer broadens, moving the dimension up a point.
▲ Product · AI Interaction Layer
AI Interaction Layer
+1 · 75 → 76
Google's Gemini Decisions API delivers calibrated multimodal decisions as a programmable endpoint, shifting the interaction layer past conversation toward a callable decision surface. Up a point.
▲ Product · AI UX Maturity
AI UX Maturity
+1 · 61 → 62
MIT's feedback-seeking design copilot and Microsoft's PDD Kit formal verification of agent-written code both advance trust and explainability at the human-agent interface. Up a point.
▼ Organization · Scaling Maturity
Scaling Maturity
-2 · 71 → 69
HFS puts only 13 percent of banks and capital markets firms at enterprise-wide scale despite spend growing near 25 percent a year, and the Gartner cancellation forecast overhangs the category. Faster pilot-to-production cycles do not offset the scaling shortfall. Down.
▼ Organization · ROI Impact
ROI Impact
-2 · 67 → 65
HFS records business-case success at 33 percent, falling to 28 percent as AI spans functions, and the MIT NANDA 95-percent no-P&L finding resurfaced. Kyndryl's 64 percent positive-ROI figure is self-reported satisfaction and does not offset the measured business-case gap. Down.
▼ Product · Agentic AI Deployment
Agentic AI Deployment
-1 · 76 → 75
Sinch reports 74 percent of organizations with live communications agents rolled back or shut down a deployment after governance failures, and Gartner expects over 40 percent of agentic projects canceled by end of 2027. Broad multi-stage deployment offsets but does not overcome production reversals. Down.
¶
The full read · 15 dimensions
Fifteen dimensions, three categories
Cell shade tracks the score; the band label carries it in words, so nothing rides on color alone.
Organization72 · Systemic
Lower
Higher · shade within each category hue
¶
Organization · Workforce AI Access
Workforce AI Access
Systemic · 73/100 · no change week over week
Access holds flat. The evidence this week confirms the baseline without extending it: 78 percent of Global 2000 companies run at least one AI workload in production, and workplace assistant spend remains a large application category. These figures restate established penetration rather than add new ground. The score holds.
¶
Organization · Scaling Maturity
Scaling Maturity
Systemic · 69/100 · -2 week over week
Scaling maturity falls. The week carried mixed evidence, and the setbacks outweigh the gains. Kyndryl records experimentation dropping below half as firms pivot to production, and pilot-to-production time has compressed sharply. Against that, HFS puts only 13 percent of banks and capital markets firms at enterprise-wide scale even as spend grows near 25 percent a year, and Gartner's canceled-project forecast sits over the whole category. The Compiled Corporation thesis depends on decision-making moving into production systems. This week the data says spend is scaling faster than the operating model behind it. The dimension moves down.
¶
Organization · Governance & Ethics
Governance & Ethics
Systemic · 79/100 · no change week over week
Governance holds. This is the Identity Control Surface dimension, and the week delivered evidence in both directions that nets flat. On the structural side, New York's RAISE Act opened its registration portal with real penalties, and governance frameworks increasingly specify named owners per system. Against that, 96 percent of enterprises report significant governance gaps and only 16.9 percent of governance measures have an explicit owner. The regulatory floor is rising while enterprise practice lags it. The score holds at a systemic level that the practice has not yet earned on the ground.
¶
Organization · Talent & Upskilling
Talent & Upskilling
Systemic · 72/100 · no change week over week
Talent holds. Leadership structure is forming faster than fluency: the CAIO seat is being filled at nearly four times the CTO turnover rate, and firms like Deloitte are attaching upskilling mandates to the role. This is ownership of the talent question, not yet evidence that fluency has broadened across the workforce. The score holds.
¶
Organization · ROI Impact
ROI Impact
Systemic · 65/100 · -2 week over week
ROI falls. The week brought the sharpest concentration of counter-evidence in the category. HFS puts business-case success at 33 percent, dropping to 28 percent when AI spans several functions, and MIT NANDA's finding that 95 percent of generative AI pilots showed no P&L impact resurfaced in fresh reporting. Kyndryl's 64 percent positive-ROI figure is the honest offset, but it is self-reported satisfaction, and even that report pairs it with falling readiness. When measured against business cases rather than sentiment, value delivery is a minority result. The dimension moves down.
¶
Brand · AI-Native Messaging
AI-Native Messaging
Operational · 47/100 · no change week over week
Messaging holds. The week's signal is tangential: vendor and analyst framing is shifting toward accountability language, a sign that AI-native messaging is maturing past capability claims. Direct evidence on enterprise brand comms fluency is thin this week. The score holds.
¶
Brand · Agent-Ready Infrastructure
Agent-Ready Infrastructure
Systemic · 63/100 · +1 week over week
Agent-ready infrastructure rises. This is the Decision Surfaces dimension, and two signals move it. ChatGPT Wallet puts a governed payment rail behind agent actions with user approval and existing card protections, placing the human-agent interface directly on the transaction. Menlo documents agent infrastructure scaling as its own spend category inside a $99.5B infrastructure pool. The machine-readable, agent-addressable layer is broadening. The dimension moves up a point.
¶
Brand · Market AI Perception
Market AI Perception
Active · 36/100 · no change week over week
Perception holds. The credibility gap stays visible: a 76 percent survey claim sits against a verified count of 24 enterprise-wide CAIOs in the Fortune 500. That space is exactly where AI-washing scrutiny lives, and it keeps this dimension pinned at the low end of operational. The score holds.
¶
Brand · Content AI-Readiness
Content AI-Readiness
Active · 37/100 · no change week over week
Content readiness holds. The relevant signal is tangential: a global search upgrade that routes more content consumption through an AI layer, which raises the stakes for machine-readable content. No direct enterprise evidence on content structuring this week. The score holds.
¶
Brand · AI-First Orchestration
AI-First Orchestration
Active · 38/100 · no change week over week
Orchestration holds. The closest evidence is adjacent: Amplitude's Wave agent runs an identify-build-measure loop, a pattern that maps to marketing and brand operations but sits in product ops. No direct evidence of brand operations run by AI this week. The score holds.
¶
Product · AI Interaction Layer
AI Interaction Layer
Systemic · 76/100 · +1 week over week
The interaction layer rises. Google's Decisions API is the material move: a calibrated, multimodal decision endpoint shifts the interaction layer past conversation and toward a programmable decision surface that applications call directly. Device expansion widens the assistant footprint alongside it. The dimension moves up a point.
¶
Product · Agentic AI Deployment
Agentic AI Deployment
Systemic · 75/100 · -1 week over week
Agentic deployment falls. The week's counter-evidence is heavy and specific. Sinch reports 74 percent of organizations with live customer communications agents rolled back or shut down a deployment after governance failures, and Gartner expects over 40 percent of agentic projects canceled by end of 2027. BeamSec's reading of broad multi-stage deployment is real and offsetting, but the rollback and cancellation data document deployments reversing in production, not expanding. Under the symmetric rule, cited setbacks that outweigh the week's positive evidence move the dimension down. They do here.
¶
Product · AI UX Maturity
AI UX Maturity
Systemic · 62/100 · +1 week over week
UX maturity rises. Two research signals move trust and explainability forward. MIT's structural-design copilot builds feedback-seeking directly into the human-agent interface, and Microsoft's PDD Kit proves agent-written code meets developer intent so reviewers check intent rather than every line. Both land on the Decision Surfaces question of where human judgment sits. The dimension moves up a point.
¶
Product · Process Redesign
Process Redesign
Systemic · 61/100 · no change week over week
Process redesign holds. The evidence nets flat. Shield shows genuine redesign, agents handling legacy-portal navigation while humans keep judgment, and that is the pattern this dimension rewards. Against it, the Superblocks and MIT Sloan field study names legacy systems and absent data pipelines as a leading reason agents fail in production, which is the signature of surface automation over real redesign. Progress and drag cancel. The score holds.
¶
Product · AI Value Delivery
AI Value Delivery
Systemic · 66/100 · no change week over week
Value delivery holds. The customer stories are real and measured: Gondola doubled revenue with a 50 percent engineer-output lift, and Holiday Extras reports higher booking conversion. These are strong anecdotes concentrated in coding-assisted development. Set against the category-level ROI evidence that only a third of enterprises meet their business case, the balance is a steady score rather than a rise. The dimension holds.
For agents · priced data cuts
This issue is free to read in full. Agents that need a specific historical or category cut on demand settle it over x402: USDC on Base mainnet, gasless, the signed data returned in the same exchange. Prices render from live route config.
GET /x402/ai-index/compositesThree composites, with history0.005 USDC
GET /x402/ai-index/organizationOrganization category · 5 dimensions0.01 USDC
GET /x402/ai-index/brandBrand category · 5 dimensions0.01 USDC
GET /x402/ai-index/productProduct category · 5 dimensions0.01 USDC
GET /x402/ai-index/dimensionAny single readiness dimension0.005 USDC
Free, always: this canonical issue, every claim endpoint, the deltas, and the methodology.
Methodology & sources
15-dimension enterprise AI readiness framework scored 0-100 across Organization (5), Brand (5), and Product (5) categories. Scores derived from public signals: news and company announcements found by web search, and the published feeds of AI vendors, research labs, and the technology press. Dimensions equally weighted within each category.
Awareness 0–20Active 21–40Operational 41–60Systemic 61–80Transformational 81–100
Produced by Applied Identities using specialized AI analysis. All scores based on publicly available data. Internal deployment data not captured. This analysis is independent and does not represent the views of any company named herein.
Produced by 3Jane Intelligence under the Applied Identities byline · methodology 1.0. Sources are cited as sources, not endorsements. The full machine payload is at /ai-index/2026-W46/payload.json.