Lovable Labs PlatformWhat is the Will AI Take My Job?
| What it answers | Automation exposure score for any job title. |
|---|---|
| How the answer is produced | Every job is a bundle of tasks, and language models are not equally good at all of them. |
| What you need to enter | Type the job title you actually do day to day, not the title on your contract. |
| Where it stops being reliable | It cannot see your employer. |
| Cost and sign-up | Free, runs in your browser, no account and no stored inputs. |
How is the automation exposure score built?
Every job is a bundle of tasks, and language models are not equally good at all of them. The score starts by mapping your job title onto a task profile: how much of the working day is spent producing routine text, moving structured data between systems, making judgement calls with incomplete information, or physically handling objects and people.
Each of those task families carries a different exposure weight. Routine drafting, summarising, first-pass data classification and template work score highest, because a current model can produce an acceptable first draft in seconds. Tasks that require accountability, physical presence, live negotiation, or responsibility for a bad outcome score lowest — not because a model cannot attempt them, but because organisations rarely hand those tasks over without a human signing off.
The weighted mix produces a 0–100 exposure number, which is then banded into a plain-language verdict. A high score does not mean the job disappears. In most exposed roles the realistic outcome is compression: the same work is done by fewer people, each of whom supervises model output instead of producing every artefact by hand.
How do you use the Will AI Take My Job?
- 1.Type the job title you actually do day to day, not the title on your contract. 'Marketing coordinator who writes all our email' scores differently from 'marketing coordinator who runs events'.
- 2.Read the band, not just the number. The difference between 61 and 66 is noise; the difference between the 'assisted' band and the 'compressed' band is not.
- 3.Look at the task breakdown and mark which lines genuinely describe your week. If half of them do not apply, your personal exposure is lower than the title-level score.
- 4.Take the two lowest-exposure tasks in your role and deliberately spend more of your week there over the next quarter. That is the practical action the score is meant to trigger.
What can this tool not tell you?
- It cannot see your employer. A well-capitalised firm with a compliance-heavy process automates far slower than a ten-person agency, even for the identical job title.
- It cannot price your relationships. Client trust, institutional knowledge and internal reputation are the most common reasons an exposed role survives a restructure.
- It is a directional model, not a forecast. Treat the output as a prompt to plan, not as evidence about your specific employment.
What should you know about reading your automation exposure number without panicking?
An exposure score is a snapshot of task composition, not a verdict on your employability. Two people with the same job title can score very differently once you account for who actually does the drafting, who owns the client relationship, and who is accountable when something goes wrong. Before acting on the number, separate the tasks that are 'produce a first draft' from the tasks that are 'decide this is good enough to send'. The former compress fast; the latter rarely do, because organisations still want a named person answerable for the outcome.
The most common misreading is treating the score as a countdown clock. It is closer to a compass: it tells you which direction the demand for your current task mix is moving, not when it arrives. A score of 70 in a slow-moving, regulated sector can take years to show up as fewer roles, while a score of 55 in a fast-moving digital agency can compress within a single budget cycle. Sector adoption speed matters as much as the raw task score, which is why this tool is best paired with attention to what is actually changing at your own employer.
What changes the answer most is granularity. Job titles hide enormous variation in day-to-day work, so the honest move is to run the score twice: once for the title on your contract, and once for the specific tasks you spend most hours on. If those two scores diverge sharply, trust the task-level one and use the gap as a guide to which parts of your role to deliberately grow.
What do worked examples look like?
A copywriter who mostly drafts product descriptions
Entering 'copywriter' returns a high exposure score because the tool defaults to the most common version of that title: high-volume, template-driven short-form text. If that genuinely describes the role, the honest response is to add a task the score does not credit — client strategy, tone-of-voice governance, or review of AI-drafted copy — before the next review cycle.
A operations manager who also does scheduling
Typing 'operations manager' returns a moderate score, split between highly exposed scheduling and reporting tasks and much less exposed tasks like vendor negotiation and incident escalation. The task breakdown shows scheduling as the single largest exposed line, which is the concrete thing to hand off to a tool rather than a person over the next few months.
What do people ask most about this tool?
Does a high exposure score mean I will lose my job?
No. It means a large share of your current tasks can already be drafted by a model. Historically that shows up first as hiring slowdowns and wider spans of control rather than layoffs. The people most affected are usually the ones hired next year, not the ones already in the seat.
Why do two similar job titles score differently?
Because the underlying task mix differs. A copywriter who produces high volumes of short-form product text has more automatable output than a copywriter who runs research interviews and brand positioning work, even though both share a title.
What is the single best thing to do with a high score?
Move up the accountability ladder. Take ownership of the review step, the client relationship, or the definition of what good output looks like. Those are the parts of the workflow that remain a human's job even after the drafting is automated.
Is the score based on my personal data?
No. The calculation runs entirely in your browser from the title you type. Nothing is stored, sent to a server, or associated with you.
Which related tools should you try next?
Written and reviewed by Jim Vernon, Editor, AI Intelligence International. Published by AI Answer Engine, a service of AI Intelligence International, and checked against our editorial standards.