How NMR uses AI responsibly
At NMR, we believe technology should support hiring, not control it.
Recent regulatory changes have placed greater scrutiny on how AI is used in recruitment, particularly around automated decision-making. Many platforms use AI to filter, rank, or reject candidates without human involvement.
NMR takes a different approach.
AI as a support tool — not a decision maker
NMR does not use AI to make hiring decisions.
Our platform:
- Does not automatically reject candidates
- Does not use AI to rank or score applicants
- Does not shortlist on behalf of employers
The match percentage you may see on a job or an application is worked out by fixed rules from the requirements the employer set, such as the role, DBS level, shifts and care setting, not by AI. It is a guide for both sides and never rejects anyone.
All hiring decisions are made by real people, ensuring fairness, accountability, and transparency throughout the process.
Enhancing efficiency without adding risk
We use AI only to simplify and speed up administrative tasks, such as:
- Assisting candidates in building and completing their profiles
- Reading an uploaded CV to fill in a profile, and wording the CVs we generate
- Suggesting a short message to send with an application, which the jobseeker can change or remove
- Drafting job descriptions from an employer's details
All of these use Claude, an AI model made by Anthropic, which doesn't use what we send to train its models. The only other service involved is Deepgram, which turns speech into text when someone uses voice input on the website. Messages between users are never sent to an AI: they are checked by fixed rules that refuse links and block abusive language and common scam patterns.
This allows both employers and candidates to save time, without removing human judgement from the process.
Andrii, our AI assistant
Jobseekers can chat with Andrii to find jobs that suit them, check what a job needs, keep their profile up to date, make a CV and get applications ready.
- Andrii asks first. Before its first answer, and before AI reads a CV, the jobseeker sees what is sent to Anthropic and chooses Allow or Not now. With Not now, nothing goes to Anthropic and they can still search and apply for jobs themselves. They can change their answer at any time in Settings.
- What Andrii sends, once allowed: the chats, the profile details it needs to find jobs, and the CV when they upload one or ask Andrii to write one or a message to an employer.
- Andrii works for the jobseeker. It doesn't rank, score or shortlist anyone for employers, and employers don't see the chat.
- Nothing reaches an employer unless the jobseeker presses Confirm: Andrii never applies, withdraws an application or answers an invitation on its own.
- Andrii runs on Claude, an AI model made by Anthropic, which doesn't use the conversations to train its models.
- Andrii can read a jobseeker's own CV to help with their applications. It never opens their other documents, such as a DBS certificate, ID or visa.
- A CV Andrii makes with a jobseeker is a draft they check before saving. A CV they upload in the chat is read by Claude to fill in their profile. On the website, spoken messages are turned into text by Deepgram, a speech-to-text service, which doesn't keep the recording.
- Andrii is an AI and can get things wrong. Every reply has a thumbs up and a thumbs down. A thumbs down with "It's offensive/inappropriate" sends that reply, the question and anything they add to our team; we look at every report within 24 hours.
- The chat is kept until the jobseeker clears it or deletes their account. Our privacy policy explains what is sent and who can read it.
Direct, transparent hiring
NMR is built to connect employers and candidates directly.
- Employers choose who they engage with
- Candidates decide where they apply
- Communication happens openly, without hidden filters or automated outcomes
There are no "black box" decisions and no unexplained rejections.
Clear criteria, no hidden bias
Employers set clear requirements for each job, such as qualifications, experience or compliance checks, and they are shown on the job. This ensures:
- No hidden filtering logic
- No algorithmic bias in decision-making
- A fair and transparent experience for all users
Data protection and user control
We take data protection seriously.
- Candidate information is securely stored
- Documents remain private and are only shared when the candidate chooses
- We do not rely on opaque third-party decision-making systems
Users remain in full control of their personal data at all times.
A human-first approach to hiring
NMR combines the efficiency of modern technology with the accountability of human decision-making.
By keeping people at the centre of the hiring process, we reduce risk, improve transparency, and create a better experience for both employers and candidates.