How our AI job score works
Last updated · aiagentsdaily.org
The free job audit gives your job a score. This page says, in plain words, where that score comes from, how to read it, and what it cannot tell you.
The short version
- Your answersYour job, and your tasks if you add them
- Task mappingAn AI model splits the job into tasks
- Hand-scored examplesFive jobs we scored keep it on scale
- Our tool listOnly live agents and guides from this site
- Your job scoreA guide, not a forecast
How we calculate the score
- Identify the tasks
You type your job. You can add your daily tasks, your industry and your level. An AI model splits the job into 4 to 6 everyday tasks.
- Map what AI can do
Each task gets a number from 0 to 100: how much of it AI can do or help with today.
- Sort the impact
Tasks with a high number are the ones AI can speed up most. A separate list shows what stays human: judgment, trust, taste and hands-on work.
- Check against examples
Five example jobs we scored by hand keep the model's numbers in line, so a new job is scored on the same scale.
- Show a balanced result
You get one overall score, the task bars, what stays human, and next steps. Suggested agents and guides are picked from this site's own list of live ones.
If the live audit is down, a built-in word list gives a rough score instead, and the report says "Offline estimate". The "What next" tab runs one live web search and links to the pages it read.
How to read your result
- The big score is the share of your usual work that today's AI can do or help with. It is not the chance you lose your job.
- The task bars show each task from 0 to 100. Green bars are tasks AI helps with a little. Red bars are tasks AI can do most of.
- The risk band is one word for the whole job: Low, Low-Medium, Medium, Medium-High or High. The offline estimate uses only Low, Medium and High.
- What stays human lists the parts of your job that still need you. A high number there is good news.
Research we read
These studies shape how we think about jobs and AI. The score does not pull numbers from these datasets.
- O*NET 31.0 DatabaseO*NET (U.S. Department of Labor)
Describes each occupation's tasks, work activities and skills. Updated every quarter.
- OECD Employment Outlook 2023: Artificial Intelligence and the Labour MarketOECD
Counting all automation technologies, AI included, 27% of jobs are in occupations at high risk of automation.
- Generative AI and Jobs: A Refined Global Index of Occupational Exposure (Working Paper 140, 2025)ILO
One in four workers worldwide is in a job with some generative AI exposure. Because most jobs still need people, the likeliest change is that jobs transform, not vanish.
- The Future of Jobs Report 2025World Economic Forum
Employers expect 170 million jobs to be created and 92 million displaced by 2030, a net gain of 78 million.
Every one of our job pages names the research it draws on, and 5 of the 20 link their sources. Their scores are hand-set examples, not audit results.
What the score cannot tell you
- AI changes fast. A score from today can be out of date in a few months.
- The same job differs from company to company, industry to industry, and place to place.
- It is an AI estimate, so two runs for the same job can differ a little.
- The score is a guide. It is not a prediction of layoffs, and it is not a guarantee.
Updated and reviewed
- The audit method last changed on .
- Each audit result is made by the AI model when you ask. No person reviews it before you see it.
- 5 of 20 job pages show a "Last reviewed" date. The latest is .
Change log
- This page published.
- The audit now picks the agents and guides it suggests from this site's own list of live ones. Before, that list held made-up examples with no page.