The State of AI Workforce Readiness in America: 144 Million Jobs, Scored

America is having its AI workforce debate with almost no instruments on the table. Executives announce AI reorganizations, workers read exposure headlines, and states fund retraining programs, and nearly all of it runs on anecdote. So we measured it. JobRoute has now scored the AI readiness of every US state, the District of Columbia, and the nation as a whole: 144,555,948 jobs, every occupation in the O*NET taxonomy, on current public data. The national AI Ready Score is 63.7 out of 100. This report walks through what that number means, which occupations and states carry the exposure, and what is already visible in payroll data. All of it is public, downloadable, and checkable.

What we measured, and what a 63.7 means

The AI Ready Score is a 0 to 100 composite computed for every occupation in the O*NET 30.3 taxonomy and rolled up by real employment weight. It combines four pillars: task exposure (how much of the occupation’s task mix current AI tools can perform, grounded in observed usage from the Anthropic Economic Index rather than speculation), skill durability, role trajectory (ten-year BLS outlook), and adjacency breadth (how many realistic role moves the occupation’s skills support). Employment weights come from BLS Occupational Employment and Wage Statistics, State, May 2025, the newest available vintage.

A national 63.7 is neither alarm nor comfort. It says the average American job retains meaningful durable skill against current AI capability, while a well-defined set of occupations, employing millions, carries task exposure high enough that the work will be substantially reshaped within a planning horizon. The score is a map of where change concentrates, not a countdown clock.

The ten most exposed occupations in America

By share of current tasks AI tools can already perform, weighted across 144.6 million jobs:

Rank	Occupation	Task exposure	Workers
1	Customer Service Representatives	84	2,595,760
2	Computer Programmers	84	91,270
3	Data Warehousing Specialists	84	33,345
4	Market Research Analysts and Marketing Specialists	83	449,790
5	Online Merchants	82	214,880
6	Business Intelligence Analysts	81	87,468
7	Medical Records Specialists	80	194,040
8	Wholesale and Manufacturing Sales Representatives	79	1,238,180
9	Medical Transcriptionists	76	39,020
10	Mail Clerks and Mail Machine Operators	76	55,220

Two observations. First, this is not a list of low-skill work. Programmers, analysts, and marketing specialists sit beside mail clerks. Exposure follows the shape of tasks: language, pattern, retrieval, and structured judgment on a screen. It does not follow the prestige of the title.

Second, scale matters more than rank. Customer Service Representatives alone outnumber the next six occupations on the list combined. When a single 84-exposure occupation employs 2.6 million people, the national conversation about AI and work is, to a first approximation, a conversation about service and back-office work.

The state rankings, and the knowledge-economy paradox

The full ranking of all 51 jurisdictions is public. The top and bottom of the table:

Rank	State	Score	Workers
1	Wyoming	65.4	259,621
2	Alabama	64.8	1,956,275
3	Alaska	64.8	299,659
4	Hawaii	64.8	586,162
5	North Dakota	64.6	397,512
47	Florida	63.5	9,483,992
48	Maryland	63.2	2,602,538
49	New Jersey	63.2	3,983,569
50	New York	63.1	8,664,786
51	District of Columbia	61.2	660,767

The spread between first and last is 4.2 points, which tells you geography matters less than occupation: what you do shapes your exposure far more than where you do it. But the ordering is the story. The most sophisticated labor markets in the country, New York, California (45th at 63.5), New Jersey, Maryland, and above all Washington DC, cluster at the bottom, while rural and resource states hold the top.

There is no mystery once you look at task mixes. AI exposure concentrates in cognitive, screen-based, information work: analysis, administration, communication, coordination. That is precisely the work knowledge economies are made of. Wyoming’s employment skews toward energy, trades, transport, and land, work that current AI tools barely touch. DC, a workforce built almost entirely of information work, sits last with the deepest pipeline decline in the country.

To be precise about what this is not: it is not a prediction that New York suffers and Wyoming wins. High-exposure regions are also positioned to capture AI productivity gains first, and PwC’s 2026 AI Jobs Barometer found the most AI-exposed firms grew both headcount and wages faster than the least exposed. Exposure measures how much of today’s task mix is in motion. In the big knowledge states, a lot of it is.

It is already in the payroll data

Employment in the most AI-exposed roles declined 1.9 percent year over year in our national pipeline measure. The state pattern matches the exposure pattern: New York’s exposed-role employment fell 2.2 percent and DC’s 2.5 percent, against 1.3 percent in Wyoming. This is consistent with what the Stanford Digital Economy Lab and ADP Research documented in payroll records, where employment for early-career workers in the most exposed occupations has measurably declined.

Hold that beside a second number. Gartner reviewed announced layoffs from the first half of 2025 and found less than 1 percent were actually caused by AI productivity gains, and it forecasts that by 2027 half of the organizations that attributed cuts to AI will be rehiring for those roles. Both things are true at once: the task shift is real and visible at the margin, and the corporate narrative is running well ahead of the measured returns. That combination, real change plus unmeasured decisions, is the strongest argument we know for instrumentation.

What to do with this, by seat

If you work in an exposed occupation: treat the score as a planning signal, not a verdict. The realistic move is a short, skill-adjacent hop, not reinvention. You can check your own role free in about three minutes at ready.jobroute.ai, including which of your tasks carry the exposure and which adjacent roles your skills already reach.

If you lead workforce strategy at an employer: the national table above is your industry’s shape at low resolution. The version that matters to you is your own org chart: which roles, which cohorts, what the retraining pathway costs against replacement. That is what JobRoute Companies does on your own HRIS data, de-identified, on the same public, auditable methodology as this report.

If you serve a state or workforce board: your state’s full overview, occupation table, and open data are already published at gov.jobroute.ai, free, with no login. If your team wants the full assessment presented, request a briefing from any state page.

Check our work

Every number in this report traces to a public source: BLS OEWS State May 2025 for employment, O*NET 30.3 for occupational tasks, the Anthropic Economic Index for observed AI usage, Stanford Digital Economy Lab / ADP Research for the pipeline measure. The methodology is versioned and published, the state data is downloadable as CSV and JSON, and JobRoute is a private company doing independent analysis of public data, affiliated with no government agency. If you find a number you disagree with, the inputs are all there. That is the point of building in the open: a score you cannot check is just an opinion with a decimal point.