Terms of Service

These terms, and who they bind

An agreement between you (the organisation using Kandevo AI) and Kandevo Limited. By creating a workspace you accept them.

If you are an individual practising on yourself, or a candidate holding a candidate account, these are not your terms: the Candidate Terms at /candidate-terms are, and nothing here binds you.

Kandevo Limited is a company registered in England and Wales, number 17427689, with its registered office at C/O GP Fund Solutions UK Limited, Mocatta House, Trafalgar Place, Brighton, BN1 4DU. Notices can be sent there or to privacy@kandevo.ai.

If you are accepting on behalf of a company, you are confirming you have authority to do so.

What the service does — and what it does not

Kandevo AI generates live interview material from a job description, records evidence against a rubric, runs simulations, and drafts summaries.

It does not make hiring decisions, and you must not use it as though it does. Every advancement, rejection and offer is yours. AI-assisted recommendations, where available, are advisory, off by default, and clearly labelled. You remain the decision-maker and the employer.

We generate interview material with a large language model. Model output can be wrong, oddly weighted, or unsuitable for a particular role. Read it before you use it. You are responsible for the questions you actually ask and the judgements you actually make.

What you are agreeing about the AI

This is the paragraph the signup checkbox refers to, so it is written plainly.

The AI is not perfect and will sometimes be wrong. It can misread an answer, weight something oddly, miss what mattered, invent a detail, or produce a summary that reads confidently and is not accurate. That is a property of the technology, not a defect we expect to remove.

You must check its work. Read the evidence behind a score before you rely on the score. Read a drafted summary before you forward it. The product links every number to the moment that produced it precisely so this is possible in seconds — but doing it is your job, not ours.

A human makes every decision here, and that human is you. Nothing in this product advances, rejects or ranks a candidate on its own.

You agree that we are not liable for mistakes or errors in AI-generated output, or for any decision, action or omission based on it. This does not affect the carve-outs in the Liability section below, which apply regardless and cannot be signed away.

Your obligations around candidates

These are the terms most likely to matter if something goes wrong, so they are not buried.

• Tell candidates. You must inform every candidate that an AI-assisted interview tool is being used, before the interview. This is a condition of using the service, not a suggestion.
• You are the controller. For candidate data you decide the purpose and the means; we act only on your documented instructions. You need your own lawful basis and your own privacy notice covering it.
• Our obligations as your processor are set out in the Data Processing Agreement, which forms part of these terms and applies automatically to every workspace on every plan, including Free. It is the contract UK GDPR Article 28 requires, and it covers confidentiality, security, sub-processors, assistance with candidate rights, breach notification within 48 hours, deletion at the end, and your right to audit. Nothing needs to be signed.
• Do not upload special category data. Do not enter health, disability, ethnicity, religion or similar information into any field, including free-text evidence notes. The product has no place for it and we do not want it.
• Keep it lawful. You are responsible for compliance with employment and equality law in your jurisdiction — in the UK, the Equality Act 2010. We provide tooling; we do not provide legal cover for how you hire.
• Accommodate candidates. Simulations are timed and text-based. If a candidate needs an adjustment, make it. Nothing in the product prevents you conducting the interview differently.

Accounts and acceptable use

You must not:

• Share credentials, or let anyone outside your organisation use your workspace.
• Use the service to build a competing product, or to extract our prompts or generated content in bulk for that purpose.
• Attempt to bypass rate limits, quotas, or the authentication on any endpoint.
• Submit content you have no right to submit, including confidential job descriptions belonging to someone else.
• Use the free public surfaces (the demo, the daily challenge, Practice Mode) as a general-purpose AI service. They are product demonstrations and are rate-limited accordingly.

Plans, payment and quotas

Plans, prices and monthly interview allowances are published on the pricing page. Paid plans are billed in advance, monthly or yearly as you choose at checkout.

Paid plans are bought through Stripe Checkout from the pricing page. Card details are entered on Stripe's own page and never reach our servers. The contract starts, and your plan changes, when Stripe confirms the payment. Enterprise is arranged directly with us.

Everything in the cancellation section below applies to every paid plan from the moment you buy it.

Exceeding your monthly interview allowance stops new interviews until the next period or an upgrade. It never silently bills you more.

We may change prices with 30 days' notice to existing customers. Your current billing period is unaffected.

Cancellation and what happens to your data

You can cancel at any time, effective at the end of the current billing period. We do not refund part-months unless the law requires it.

Cancelling stops the billing and drops you to the Free plan; your workspace and its records stay where they are, so nothing is lost while you decide. Export and deletion are both yours to run, whenever you want. "Export everything" downloads one file holding every kit, interview, score, observation, integrity signal, screening answer and the audit trail. It is the assessment record, not our internal billing counters. "Delete this workspace" removes all of it permanently, including the audit trail and your account; it needs your password and cannot be undone. We do not delete anything on a timer, and we would rather say so than describe a countdown that does not run.

We may suspend or terminate a workspace for a material breach of these terms, non-payment, or use that puts candidates at risk. Except where the breach is serious enough to require immediate action, we will tell you first and give you a chance to fix it.

Your data and ours

You own what you put in. Your job descriptions, your interview records, your evidence and your decisions remain yours. We claim no ownership and will not use them to train models.

We own the product. The software, prompts, design and the Kandevo name are ours.

Generated kits are yours to use. Interview material generated for your role is yours to use in your hiring, without restriction. We do not claim copyright in it and we do not reuse your specific kits for other customers.

We may use anonymised signals to improve the product — that a kit was rated poorly for a role family, for instance. These carry no candidate identity and no demographic field, and a database constraint enforces it.

Availability

We aim to keep the service available and will give notice of planned maintenance where we can.

There is no uptime SLA on any published plan. Enterprise agreements can include one; nothing else does, and we would rather say that here than imply a guarantee we have not committed to.

Liability

Nothing in these terms limits liability for death or personal injury caused by negligence, for fraud, or for anything else that cannot lawfully be limited. That carve-out is required by law and comes first for a reason.

Subject to that, our total liability in any 12-month period is limited to the fees you paid us in that period.

We are not liable for indirect or consequential loss, lost profits, or lost opportunity.

We are not liable for your hiring decisions. You make them, using your judgement, with a tool that says throughout that it does not decide. A claim that our software rejected a candidate is a claim about something the product does not do.

We are not liable for mistakes in AI-generated output, or for decisions taken on it — see "What you are agreeing about the AI" above. Read it before you rely on it.

Third parties

Only you and we can enforce these terms. Nobody who is not a party to them has rights under the Contracts (Rights of Third Parties) Act 1999.

That is worth stating plainly because these terms talk about candidates a great deal — what you must tell them, what you must not upload about them, what we do with their answers. Those obligations are real and we take them seriously, but they are obligations you owe, enforceable by us, not a contract between us and the people you interview.

It does not take any right away from a candidate. Their rights come from data protection law, not from this agreement, and nothing here limits them. A candidate who thinks their data has been mishandled should contact the company that interviewed them, and may complain to the ICO — see the Privacy Policy.

Governing law

These terms are governed by the law of England and Wales, and the courts of England and Wales have exclusive jurisdiction.

Changes to these terms

This is version 1.4, effective 2026-09-19.

We will give account holders at least 30 days' notice by email of any change that materially reduces your rights or increases your obligations.

Questions: privacy@kandevo.ai.