AI is part of how we hire and part of what we hire for. This policy sets out how LemFi uses AI in our process, where we want you to use it, and the one place where it crosses a line: using AI to invent or stand in for your own experience.

In short

We use AI to work faster, but humans make every hiring decision.
AI capability is a skill we assess for every role, scored against a clear framework like any other competency.
Use AI freely for research, case studies, presentations and take-home tasks. Be ready to explain how you checked it and why you stand behind it.
In live problem-solving interviews, we will tell you up front if AI is allowed and how.
Do not use AI to falsify your experience or to answer questions in real time on your behalf. That is the line we draw, and crossing it can end your application.
If you need an adjustment, including assistive technology, tell us and we will work it out with you.

How LemFi uses AI in hiring

We use AI to remove slow, repetitive work so our recruiters and hiring managers can spend their time talking to you. AI supports our decisions; it never makes them. Every decision to reject, progress or hire you is made by a person who can see why the tool said what it said, and who can overrule it.

Where	What AI does	What a person does
Applications	Scores applications against criteria set for each role, so the most relevant ones are reviewed first	A recruiter reviews applications before anyone is rejected
Fraud checks	Flags signals that an application may be fake, duplicated or misrepresented	A recruiter investigates every flag. A flag alone never rejects you
Interviews	Ashby's AI notetaker records and transcribes the conversation so your interviewer can focus on you	Interviewers score you against the role's criteria and write their own feedback
Feedback review	Summarises feedback across interview stages to spot gaps and inconsistencies	The hiring team discusses and decides
Sourcing	Juicebox and our sourcing agents search public professional profiles and help draft outreach	A recruiter decides who to contact and personalises the message
Internal tools	Our own tools draft job descriptions, interview packs and other repeatable documents	The hiring team reviews and signs off everything before it is used

What we will not do

Use AI to analyse your face, voice, tone or body language, or to infer your emotions or personality.
Use AI text detectors to judge whether you wrote something yourself. They are unreliable and unfair to people writing in a second language.
Let any tool make a final hire or no-hire decision.

If we found you through sourcing, our first message will tell you where we found your details and link to our candidate privacy notice. You can ask us to delete them at any time and we will not contact you again. To make sure of that, we keep only the minimum details needed to recognise you.

The notetaker is your choice. We will tell you before the start of every interview that it is joining and check you are happy for it to record. If you would rather we did not record, you have the option to remove it from all calls. It will not affect how you are assessed.

AI is a skill we assess

We assess AI capability in every hiring process, in the same way we assess communication, ownership or technical depth. It is a scored competency with a defined bar for each role, not a gut feel.

We use an internal levelling framework that describes what good AI use looks like at each level, from confident everyday use through to building tools and systems others rely on. Every role has a minimum level, set by what the role needs. Most roles need you to be a capable, critical everyday user. Roles that improve how teams work, and all engineering roles, have a higher bar.

What we score

How AI has changed the way you actually work, with real examples and outcomes.
How you prompt, iterate and choose the right tool for the job.
How you check output before you rely on it, and where you have caught it being wrong.
For more senior or technical roles, what you have built that other people use.

What we do not score

How enthusiastic you are about AI. Well-evidenced scepticism is a strength, not a red flag.
Which paid tools you have access to. Where a task needs AI, free tools are fine, or we will provide access.
Jargon. We care what you did and why, not the vocabulary.

Your recruiter will tell you which stages assess AI capability and what level the role requires before you reach them.

Where we want you to use AI

Use AI wherever it makes your work better and faster. We would rather see how you work with it than pretend it does not exist.

Your application. Use AI to structure and polish your CV or cover letter. Everything in it must be true and yours. If it says you did something, expect to be asked about it in detail.

Research and preparation. Use it to understand LemFi, our markets and the role. Good preparation shows.

Case studies, presentations and take-home tasks. Use it freely for research, analysis, drafting and building slides. Because AI makes a competent first draft easy, the bar is higher. We are not grading what any model could produce. We are grading your judgement: what you chose to keep, what you challenged, and how you made it specific to LemFi. Expect us to ask:

Which tools you used, and why.
How you checked that the facts, figures and sources were right.
What the AI got wrong, and what you changed.
Why you believe your recommendation, in your own words, and what you would do if we pushed back on it.

If you cannot explain or defend something you presented, it will count against you, whoever or whatever wrote it. Do not put anything we have marked as confidential into a public AI tool.

Live problem-solving interviews. Some interviews, especially for engineering, product and data roles, will ask you to use AI tools live. When that happens, we will tell you in your interview preparation which tools are allowed and what we are assessing. If we have not said AI is allowed in a live session, assume it is not. If you are unsure, ask. Asking will never count against you.

Where AI use is not acceptable

Your experience has to be yours. When we ask about what you have done, how you handled a situation or what you learned, we are assessing you, not a model. Using AI to answer those questions, or to invent or inflate your experience, is unacceptable at every stage of our process.

This includes

Real-time assistants such as Cluely, or any LLM, generating or feeding you answers during an interview, unless we have said that tool is allowed in that session.
Reading out AI-written answers to questions about your own experience.
Fabricating or exaggerating roles, results or examples, in your application or in conversation, whether AI wrote them or you did.
Anyone or anything other than you taking part in an interview, including AI-generated video or voice.
Using AI in a live session where we have not said it is allowed.

What is fine: using AI before an interview to practise, run mock questions, or help you recall and structure your real examples. Preparation is not the issue. Substitution is.

What happens if we see it

We may end your application at any stage, and may withdraw an offer if we find out after it is made. We do not act on a hunch, and we do not use AI detectors. We act when the evidence is clear, for example what we see and hear during the interview. We score answers that are vague or fall apart under follow-up questions on their merits, like any other answer, but on their own we do not treat them as misuse.

When we have a concern, we will usually raise it with you and ask more follow-up questions, change the format, or invite you to a further conversation so you can show us your experience directly. A decision to reject you or withdraw an offer on these grounds is reviewed by at least two people, including a senior member of the Talent team, before it is made. If you think we got it wrong, tell us and we will look again.

Adjustments and your rights

Adjustments: If you have a disability, health condition or are neurodivergent, and you use assistive technology such as speech-to-text, live captioning, a screen reader or writing support, tell your recruiter, ideally before your interview. We will agree with you on what is appropriate. Agreed assistive tools are never treated as misuse. You can ask for other adjustments too, such as questions in advance, extra time or a different format.

Your rights: At any point you can:

Ask how AI was used in your application and what role it played.
Ask for a person to review any decision about you, and challenge it.
Ask us not to record your interview.

Email careers@lemfi.com and we will respond as quickly as possible. Our candidate privacy notice explains how we handle your data in full. If you are unhappy with our response, you can raise a complaint with your local data protection authority.

We review this policy regularly, because the tools, and the rules, are moving fast.