Privacy and terms
Last updated 7 October 2026. This is a free service. The notice below describes what the app actually does today; when the product changes, this page changes with it.
Who runs this
AI Overlap Score is run by its individual operator (the data controller). Contact for anything on this page, including privacy requests: luisroguzgom@gmail.com. We aim to answer within one month.
What this service is
AI Overlap Score produces an automated, AI-generated reading of how exposed your job tasks are to current AI tools. Results are generated by a machine learning model, are advisory only, and can be wrong. They are shown only to you; no decision about you is made, and nothing about you is shared with employers or anyone else.
Running an analysis (no sign-in needed)
To run an analysis, the app sends the job title, industry, seniority and the 3 to 8 task descriptions you typed or confirmed. That is the whole payload. Do not include your name, employer names, contact details, or anything sensitive in task text; the form does not need it and the analysis does not use it.
If you use the CV import: the PDF is read entirely in your browser and is never uploaded. The app proposes task lines from it, strips obvious identifiers (emails, phone numbers, links), and shows you everything before anything is sent. Only the task lines you confirm leave your device.
The optional capacity calculator on the results page runs entirely in your browser. Nothing you type into it is sent to us or stored. We do not ask for salary anywhere.
Unless you sign in and choose to save, nothing you submit is stored by us after the response is returned, beyond a short-lived in-memory cache of identical requests. Our server keeps operational logs of each request (timing, whether it succeeded, and the failure reason if not); these logs never contain your task text or job title.
Signing in and saving (optional)
You can sign in with a Microsoft or GitHub account through Azure Static Web Apps' built-in authentication. We never see your password. From the provider we receive an opaque user identifier, the provider name, and the display name or email the provider supplies. We store the identifier and provider name to find your profile, and show the display name back to you; we do not use it to contact you.
If you choose “Save to my profile” on a result page, we store under your identifier: your job title, industry, seniority, the task descriptions you confirmed, the reading itself (levels, rationales, learning paths), a history of your scores and task levels with dates, and which learning steps you have ticked. This lets you come back, tick off steps, and re-run the reading (at most once a day) to see what changed. Nothing is saved unless you press that button.
Saved data lives in Azure Cosmos DB in Microsoft's East US 2 region, with Microsoft as our processor under the same data protection addendum as the analysis itself. It is kept until you delete it; a profile untouched for 12 months is deleted by us. The “Delete my saved data” button on your profile removes everything linked to your identifier immediately. To also revoke the sign-in consent held by the platform, visit /.auth/purge/aad or /.auth/purge/github after deleting.
The anonymous roles overview (opt-in)
Separately from saving, you can tick a box to contribute an anonymous summary of a reading to the public roles overview. A contribution contains only: a role bucket derived from your job title, an industry bucket, your seniority band, the overall score, the count of tasks at each level, the month, and the scoring rubric version. It contains no task text, no identifier, no link to your profile, and no timestamp finer than a month. The overview shows a role only once at least five readings have been contributed to it. Because nothing in a contribution identifies you, a contribution cannot be withdrawn after the fact; unticking the box stops future ones. Readings produced in mock mode are never contributed.
Where your data goes
Analysis requests are processed by Microsoft Azure OpenAI as our processor, under the Microsoft Products and Services Data Protection Addendum (which incorporates EU Standard Contractual Clauses). Processing occurs on Microsoft infrastructure, including in the United States. Microsoft does not use your data to train AI models. For abuse monitoring, Microsoft may retain prompts and responses for a limited period (per Microsoft's published policy) and authorised Microsoft personnel may review flagged content; we cannot access or delete that store on demand.
Automated processing
The assessment is fully automated and advisory. The model rates each task you entered against its knowledge of current AI capabilities and states its confidence; the overall score is then fixed arithmetic on those task ratings (each task weighs the same), so a rerun on the same tasks is comparable. High exposure ratings must name a specific capability, and unnamed high ratings are automatically downgraded. Task text that tries to instruct the assessor is ignored and marked as such. The capability references currently come from the model's own knowledge and are labelled as unverified. If you want to contest a result, ask how it was produced, or reach a human, email the address above.
Your rights
Depending on where you live, you may have rights to access, correct, delete, or object to processing of your personal data, and to complain to a supervisory authority. If you never saved, we hold nothing. If you did, your profile page shows everything we hold and lets you delete it; email us for anything else, including anything held by our processors.
Intended audience
At present, AI Overlap Score is offered to users outside the European Economic Area and the United Kingdom and is not directed at EEA or UK residents. If that changes, this notice and our processing arrangements will be updated first.
Acceptable use
AI Overlap Score is a self-assessment tool for individuals about their own role. It must not be used by or on behalf of employers, recruiters, or staffing platforms for recruitment, selection, evaluation, promotion, termination, task-allocation or any other workforce decision about another person. That applies to the roles overview too: it is aggregate information about how AI overlaps with kinds of work, not about any person. Automated scraping or bulk querying is not permitted.
No guarantees
The service is provided as-is, free and non-commercial, with no warranty of accuracy or availability. Exposure readings are information, not career, financial or legal advice.