The short version
- Read the site, use the ideas, quote us with a link. It is a marketing site, not a product with an account.
- Nothing here is a binding offer. Prices and timelines on the site are indicative; the agreement we sign is what governs the work.
- You own what we build for you once you have paid for it. We keep the generic tools and libraries we brought with us.
- AI systems are probabilistic. They are wrong sometimes. We will tell you where that risk sits and design around it, but we will not promise an output is always correct, and you should not build a process that assumes it is.
- Alberta law, Alberta courts.
What these terms cover
These terms are between you and Clovai Inc. of Edmonton, Alberta, Canada. They cover two different things, and it is worth being clear which is which:
- Using this website. These terms apply the moment you browse it.
- Working with us. Sections Enquiries onward are the baseline for a Clovai engagement.
Using this website
Use it as you would expect to use any publicly readable website. We ask only that you do not:
- attempt to break, overload or gain unauthorised access to the site or the systems behind it
- submit the contact form automatically, in bulk, or with someone else's details
- use the site to send anything unlawful, or anything you have no right to send us
- misrepresent yourself as connected to Clovai Inc. when you are not
You are welcome to link to us, quote us with attribution, and read the site with an automated tool for ordinary purposes such as search indexing or accessibility. We may restrict access if the site is being abused, and we may take the site down for maintenance without notice. It is a website, not a service with an uptime commitment.
Our content, and yours
Ours
The text, design, code and images on this site belong to Clovai Inc. or are used with permission. The Clovai name and logo are ours. You may quote reasonable extracts of our articles with attribution and a link; please do not republish a whole piece, present our work as yours, or use our branding in a way that implies we endorse or are involved in something.
Our articles are written to be useful, not to be advice for your specific situation. Acting on an article without talking to someone about your circumstances is your call and your risk.
Yours
When you send us a message, you keep ownership of what you wrote. You give us permission to read it, store it and use it to respond to you and to run our business — nothing more. We treat the contents of an enquiry as confidential, as described under Confidentiality, and we handle the personal information in it as set out in our privacy policy.
Please do not send us confidential information you are not authorised to share, and please do not send us an unsolicited idea you expect to be paid for. If you want something protected before you tell us, ask for an NDA first and we will sign a sensible one.
Enquiries, quotes and scope
An enquiry is not a contract, and neither is our reply to it. Any price, range, timeline or capability mentioned on this site or in an early conversation is indicative: it reflects what comparable work has cost, not a committed quote for yours.
Work begins when we have both agreed a written scope that says what is being built, what it costs, what we need from you, and when. Changes to that scope after it is agreed are normal; they are handled by agreeing the change and its effect on cost and timeline in writing before we build it, rather than afterwards.
Our estimates assume you give us the access, decisions and subject matter input the scope says you will. Where that does not arrive, the timeline moves — we will tell you as soon as we see it happening rather than at the end.
Who owns what we build
On full payment, you own the deliverables we created specifically for you: the application code we wrote for your project, the configurations, the documentation, and the trained artefacts produced from your own data. We assign that ownership to you outright, and we will sign whatever reasonable paperwork your lawyers need to confirm it.
We keep ownership of the general-purpose things we brought to the job and would bring to any job: our internal libraries, templates, tooling, methods and know-how, plus anything we had before the project started. Where a deliverable has any of that embedded in it, you get a perpetual, worldwide, royalty-free licence to use, modify and sublicense it as part of that deliverable. In practice this means you can take the code, change it, host it elsewhere, and have another firm maintain it, with no permission needed from us and no licence fee.
Your data stays yours throughout. We do not acquire any ownership of it, we do not use it to train models for anyone else, and we do not put it into a public model. On request at the end of an engagement, we will return it and delete our copies.
Third-party components — open-source libraries, a model provider's API, a hosting platform — come with their own licences and terms, which apply to you as the operator of the system. We will tell you what they are before we build on them.
We would like to describe the work publicly as a case study. We will ask you first, in writing, and a no is a perfectly good answer that costs you nothing.
Working with AI systems
This section exists because it is the part people most often assume and most often get wrong, and we would rather be unambiguous about it than have the conversation after something has gone sideways.
- AI outputs are probabilistic, not deterministic. The same input can produce different output. A system that is right 98% of the time is wrong one time in fifty, and that is a property of the technology, not a defect we can patch out.
- We will not warrant that an output is accurate, complete or fit for a decision on its own. What we will do is tell you honestly where the error rate sits, design the system so a wrong answer is caught rather than acted on, and be clear about which steps need a human.
- Where an AI system makes a consequential decision — about money, employment, credit, health, safety, eligibility or legal rights — it needs human review, and that review is yours to operate. We will build the review step; we cannot be the reviewer.
- You remain the operator. Once a system is live in your business you are the one deploying it, and the obligations that come with that — telling people they are dealing with an AI where the law requires it, handling the personal information it processes, keeping a human escalation route — sit with you. We will tell you what they are and build for them.
- Model providers change things. Models are deprecated, prices change, rate limits change, and terms change, on the provider's schedule and not ours. We design to make swapping a provider possible rather than catastrophic, but we do not control it and cannot indemnify you against it.
- Third-party model terms flow through to you. If a system uses an external model, that provider's acceptable use and data handling terms apply to the data you send it. We will name the provider and link the terms before we build.
None of this is a reason not to do the work. It is the difference between an automation that saves you real money for years and one that quietly produces wrong answers nobody checks.
Fees and payment
Fees, the payment schedule and the currency are set in the written scope for your project, because they depend on it. Unless that document says otherwise:
- invoices are payable within 15 days of the invoice date
- fees are quoted exclusive of GST and any other applicable tax, which is added where it applies
- third-party costs we incur on your behalf — hosting, model API usage, licences — are passed through at cost and are yours, not ours
- we may pause work on an overdue invoice after telling you, and we may charge interest at 1.5% per month on amounts more than 30 days overdue
- amounts already invoiced for work already done are not refundable, though we will always rather fix a problem than argue about an invoice
Confidentiality
Each of us will keep the other's confidential information confidential, use it only for the engagement, and protect it at least as carefully as our own. That obligation survives the end of the engagement and lasts for three years after it, or for as long as the information remains a trade secret, whichever is longer.
It does not apply to information that is already public, that we knew before you told us, that we received legitimately from someone else, or that we are legally compelled to disclose — in which case we will tell you first, if we are allowed to.
Warranties and what we don't promise
We warrant that we will perform the work with the reasonable skill and care of a competent professional in our field, and that we have the right to grant the ownership and licences described above. If work does not meet that standard, tell us within 30 days and we will re-perform it at our cost. That is a real commitment and we intend to honour it.
Beyond that, and to the fullest extent the law allows, the website and our deliverables are provided as is. We do not warrant that the site will be uninterrupted or error-free, that any system will be free of defects, that AI output will be accurate, or that a particular business outcome, revenue figure, ranking, cost saving or return on investment will be achieved. Any figure we discuss about what a system might save or earn is an estimate based on assumptions we will state, not a promise.
Nothing here limits any warranty or right you have that cannot lawfully be excluded, including under Canadian consumer protection legislation.
Limitation of liability
To the fullest extent permitted by law, neither of us is liable to the other for indirect, incidental, special, consequential or punitive damages, or for lost profits, lost revenue, lost data, lost goodwill or business interruption, even if told such damages were possible.
Our total liability arising out of or relating to an engagement is limited to the fees you paid us for that engagement in the 12 months before the claim arose. For use of this website alone, where you have paid us nothing, our total liability is limited to CAD $100.
These limits do not apply to our fraud, our wilful misconduct, our breach of confidentiality, death or personal injury caused by our negligence, or any liability that cannot lawfully be limited. They are a deliberate allocation of risk that the fees reflect: if your project needs a higher cap, say so during scoping and we will price it.
Changes and termination
We may update these terms. The date beside these sections changes when we do, and the version in force for an engagement is the one at the date its scope was agreed — we will not retroactively change the terms of work already underway. Continuing to use the site after a change means you accept the updated terms for the site.
Either of us may end an engagement by giving the notice its written scope specifies, or 30 days if it specifies none. On termination you pay for work done up to that point, we hand over what we have built and your data, and the sections on ownership, confidentiality, warranties, liability and governing law survive.
Governing law
These terms are governed by the laws of Alberta, Canada and the federal laws of Canada that apply there. The courts of Alberta have exclusive jurisdiction, and we each submit to them — though we would both rather sort a disagreement out in a conversation, and we commit to trying that first.
If any part of these terms turns out to be unenforceable, the rest stays in force and that part is read as narrowly as it needs to be to become enforceable.
Questions about any of this: hello@clovai.ca.