TaskExposed.
DATASET v2.4 · 900+ ROLES
Get report
RESEARCH BRIEF · MAY 2026· Updated daily as models evolveNEW

AI won't replace
jobs. It replaces tasks.

The first task-level intelligence platform for the AI economy. See exactly which parts of your work are already exposed — and what compounds.

⌘ K
METHODOLOGY ANCHORED IN
O*NET 28.3
BLS OEWS
ANTHROPIC TASK-LEVEL
OPENAI GPTs RESEARCH
ACEMOGLU & RESTREPO
Live exposure scan
0%
of an average software engineer's tasks are addressable by current frontier models.
37% remain HUMAN-CRITICAL · UPDATED 18:42 UTC
LIVE TASK STREAM
15 / 12,408 TRACKED
Generate API documentation
91%
Write unit tests
88%
Triage production incidents
24%
Draft sales emails
86%
Code review feedback
58%
Negotiate vendor contract
12%
Compile quarterly report
79%
Design system architecture
28%
Translate technical spec
82%
Mentor junior teammate
11%
Forecast revenue model
64%
Schedule meetings
94%
Run user interviews
31%
Build onboarding deck
73%
Diagnose latency regression
38%
Trending this week

The professions everyone is checking.

Browse all roles
Computer & Math
Software Engineer
Moderate
// AI EXPOSURE
63%
MEDIAN PAY
$132k
WORKERS
1.8M
10Y GROWTH
+17%
Arts & Design
Copywriter
High
// AI EXPOSURE
84%
MEDIAN PAY
$64k
WORKERS
142k
10Y GROWTH
-2%
Legal
Paralegal
High
// AI EXPOSURE
79%
MEDIAN PAY
$56k
WORKERS
378k
10Y GROWTH
+4%
Healthcare
Registered Nurse
Low
// AI EXPOSURE
31%
MEDIAN PAY
$81k
WORKERS
3.1M
10Y GROWTH
+6%
Business & Finance
Financial Analyst
High
// AI EXPOSURE
76%
MEDIAN PAY
$96k
WORKERS
371k
10Y GROWTH
+9%
Business
Customer Service Rep
High
// AI EXPOSURE
81%
MEDIAN PAY
$38k
WORKERS
2.9M
10Y GROWTH
-5%
How it works

Three steps from job title to game plan.

We don't score jobs. We score the ~30 tasks inside each job — then aggregate, weight by time-spent, and project forward against current model capabilities.

01
Step 1
Search your profession
Pick from O*NET-anchored job titles, or describe what you do — we map free-text to the closest canonical role.
›Software Engineer▍
· Software Engineer63%
· SRE41%
· Eng Manager34%
02
Step 2
Analyze task exposure
Each task is graded against frontier-model capabilities and weighted by your real time allocation. The score adapts as models improve.
Write boilerplate92%
Code review64%
System design28%
03
Step 3
Adapt before everyone else
Get the durable skills to deepen, the adjacent careers to consider, and the 90-day moves that compound.
Resilient adjacent roles
Engineering Manager-29pp
AI/ML Engineer-12pp
Security Engineer-25pp
Grounded research

Built on the same datasets economists use.

We combine the U.S. Bureau of Labor Statistics task taxonomy, Anthropic's task-level capability research, OpenAI's GPT-exposure paper, and a continuously-updated capability matrix benchmarked against frontier models.

TASK TAXONOMY
O*NET 28.3
923 occupations · 19,260 tasks
U.S. DEPARTMENT OF LABOR
LABOR DATA
BLS OEWS 2025
154M workers · wage + employment
BUREAU OF LABOR STATISTICS
CAPABILITY MAPPING
Anthropic 2024
Task-level Claude performance
ARXIV:2403.XXXXX
EXPOSURE RESEARCH
OpenAI GPTs
Eloundou et al. 2023
ARXIV:2303.10130
Take the snapshot

Understand your AI exposure
before the market does.

Free for everyone. No signup needed to view a report. Premium gives you skills mapping, custom task input, and quarterly delta tracking.

ALREADY USED BY 248,310 PROFESSIONALS · 41 COUNTRIES