AI is quickly becoming part of everyday business. It can help teams analyse data, draft content, improve customer support and deliver services more efficiently. For businesses that provide AI-enabled services to customers, however, the technology is only part of the equation.

The other question is whether your customer contract is ready for it.

You may still be using a standard software-as-a-service (SaaS), services or managed-services agreement written before AI became part of your product. It may say little, or nothing, about how AI is used, what happens to customer data, who owns the output, or what happens if the technology gets something wrong. That can lead to awkward conversations when a customer starts asking questions.

Can you use AI under the contract?

Does your agreement allow you to use AI to deliver the service?

A general right to use technology or subcontractors may not be enough. Your contract may restrict third-party platforms, overseas processing, or disclosure of confidential information. Customers should know if an external AI provider processes their information.

Clear drafting builds trust. Explain how you use AI in plain language, and make sure the contract reflects any relevant third-party providers and safeguards.

What happens to customer data?

This is often the question customers care about most. If a customer uploads documents, enters prompts or shares data through your service, what are you allowed to do with that information?

Your contract should give you the rights you need to provide the service, but no more. Be clear whether customer data can be retained, de-identified, aggregated or used to improve the service or train a model. A general right to process data may not cover AI training.

The practical safeguards matter too. Your contract and data-processing terms should reflect where data is processed, which sub-processors are involved, what security commitments apply, how quickly incidents will be reported, and what happens to data at the end of the relationship.

Who owns the intellectual property?

AI can make intellectual property (IP) ownership trickier than it first appears. Your agreement should distinguish between:
  • your existing platform, models, tools, and know-how;
  • the customer's data, materials, and branding;
  • any bespoke work created for that customer; and
  • the AI-generated output delivered through the service.

As the provider, you will usually want to retain ownership of your underlying technology and the improvements you make to it. Your customer will want clear rights to use the output it has paid for. There is room for both, as long as the agreement says so clearly. But contractual rights to use AI-generated output are distinct from whether that output attracts IP ownership under applicable law.

It is also worth being realistic about what the technology can promise. Depending on the service, you may not want to guarantee that AI output will always be accurate, complete, unbiased, unique, or free from third-party rights. The contract should set sensible expectations about how the output should be checked and used.

Who carries the risk if things go wrong?

Customers expect confidentiality and agreed service standards. But you should avoid unlimited responsibility for risks you cannot fully control, particularly where customers use output without human review.

This is where it pays to take a fresh look at your liability caps, exclusions, indemnities and warranties. A customer may reasonably seek protection for a serious security incident or IP claim. But an unlimited promise that every AI output will be correct or suitable for every purpose is unlikely to be workable.

The right balance depends on the service, the data involved, the customer's intended use and the value of the deal.

A quick AI contract checklist

Before launching or selling an AI-enabled service, ask:
  1. Does the customer agreement clearly allow the AI use and any relevant third-party providers?
  2. What customer data can be processed, kept, de-identified, aggregated or used to improve the service?
  3. Who owns the platform and the customer's input, what rights does each party need to use the output, and can the output attract IP ownership under applicable law?
  4. Do the confidentiality, security, data-processing, subprocessor and deletion provisions match the real workflow?
  5. Does the contract set realistic expectations about output accuracy, human review, and appropriate use?
  6. Do the warranties, indemnities and liability provisions fairly deal with data, IP, security, and reliance risks?
  7. Is there a clear process for material changes to the AI model, the data practices, or the service?

Make sure the contract keeps up

AI can be a competitive advantage, but your contract needs to keep up. Clear terms protect your technology, reassure customers, and reduce disputes.

Providing an AI-enabled service? Ethiqs Legal can review and update your customer contracts to clarify data and IP rights, manage AI-related risk, and help make sure your terms keep up with your product. Book a free 20-minute call with the Ethiqs team.

Book a free 20min Chat

Feel free to ask for details, don't save any questions!

Our Office

Business Hours

  • Monday - Friday - 9am to 5pm

Get in Touch

Ethiqs is committed to providing our clients with accessible, transparent and affordable legal services and this starts all the way from the initial consultation.