These terms explain our general approach. Each client's written quote or service agreement should identify the specific services, fees, data connections and responsibilities that apply to their setup. The website privacy notice separately explains information submitted through this public site.
01 / Working together
No long-term contracts
We believe clients should continue working with Owner Insights AI because the service is useful to them, not because they are locked into a long-term commitment.
- Ongoing services are provided month to month. Either you or Owner Insights AI may end an ongoing service by sending written notice by email to the contact specified in your quote or service agreement.
- You can give notice at any time; service will end at the close of the current paid monthly period. We recommend giving at least 10 days' notice before your next billing date to help avoid an unintended charge.
- Fees for a monthly period that has already begun are not refunded in part, except where required by law or agreed otherwise in writing.
One-time implementation and any optional work to correct or prepare financial records will be scoped and quoted separately. Your quote will identify the services and fees that apply to you.
02 / Answers and evidence
How we test accuracy
We configure and test each client's system against supplied records before releasing it for use. In the practical tests we have performed, the answers checked have matched the records and calculations used for those tests. This gives us confidence in the tested setup, but it cannot prove that every future answer to every possible question will be correct.
AI can produce a confident answer that is incomplete, inaccurate or based on a misunderstanding of the question. Results also depend on the quality and completeness of the underlying business records; testing an answer against those records is not an audit of the records themselves. Please check figures that will drive important financial, operational or legal decisions. Tell us if an answer appears wrong so we can investigate and correct the setup where needed. Read more about our testing process.
03 / Information entrusted to us
Client data and security
Before a client begins using the service, we agree where their information will be held, which services may process it and who needs access. We aim to limit the information shared to what is needed for the agreed service and use reasonable safeguards appropriate to that information. Access may include approved client users and designated Owner Insights AI personnel who need it for setup or support. No internet-connected system or AI service can be guaranteed completely secure.
Financial information sent to an AI provider is processed to answer requests, even if that provider does not use it to train models. Depending on the chosen setup, other providers, such as a connected app or hosting service, may also process relevant information under their own policies. We will identify the data flow and applicable storage practices for the particular service before release. Do not assume that information permitted for model training is completely anonymized or that a model could never reproduce a specific detail.
04 / ChatGPT account option
If you use Owner Insights AI in ChatGPT
Information retrieved through a connected app or plugin may become part of a ChatGPT conversation and be processed by OpenAI. Depending on the connection, the app provider may also process information needed to provide the service. Under OpenAI's current published policy, information accessed through apps in ChatGPT Business and Enterprise is not used to train its models by default. On personal plans, including Plus and Pro, it may be used for training when Improve the model for everyone is on.
Recommended account settings
For sensitive financial information, we recommend a company-managed ChatGPT Business or Enterprise workspace. If you choose a personal account, turn off Settings → Data controls → Improve the model for everyone before connecting financial information, and confirm the choice for each person who will use the service. Enable multifactor authentication, review the connection's requested access, restrict it to authorized users, and review Memory and conversation-sharing settings.
Avoid submitting feedback on conversations containing sensitive information: OpenAI says the conversation associated with feedback may be used for training even after opting out. Turning off model improvement affects new conversations; it does not erase saved conversations or stop processing needed to answer a request. App permissions, Memory and training are separate controls. See OpenAI's current data-control instructions for the settings available to your account.
05 / Web interface option
If you use a tailored web interface
A web interface sends the information needed for a request from our application to the OpenAI API. Under OpenAI's current API policy, API data is not used to train its models unless the API customer explicitly opts in. We will not knowingly opt in to model training with a client's financial information without the client's prior written authorization. OpenAI may still retain some request information for abuse monitoring, and retention varies with the API features used. Its default abuse-monitoring logs can contain request content and are generally retained for up to 30 days, subject to the exceptions described in its policy.
An API-based design gives us more control over what our application sends, which client users can access it and what our application stores. Its overall security depends on the safeguards actually implemented across the application, hosting and service providers. Both the ChatGPT and web-interface options involve processing by OpenAI; neither offers an absolute security guarantee.
06 / An informed choice
Need help deciding?
You can share this public page with an AI assistant you trust and ask it to explain the advantages, limitations and setup steps for the options described here. Tell it which type of ChatGPT account you use and how you prefer to work, but do not provide financial statements, passwords or other confidential information just to get that guidance. Ask it to link to the current official instructions for any setting it recommends, then check the settings in your own account before connecting business information. If your organization has an IT or security adviser, involve them in decisions about sensitive data.
07 / Reasonable limits
Limits of responsibility
We test each client's setup before release and take reasonable steps to protect information entrusted to us. AI answers can still be incorrect or misunderstood, and no connected system can be guaranteed free from security incidents. Please review important answers against your source records and use your own judgment, and professional advice where appropriate, before making financial or operational decisions.
To the extent permitted by applicable law, Owner Insights AI is not responsible for indirect or consequential losses, such as lost profits or business opportunities, arising from an incorrect or misunderstood AI answer, a client's account settings or access decisions, or a security incident outside our reasonable control. This does not remove any responsibility that the law does not allow us to exclude or our obligation to meet the commitments we make to clients.
Last updated September 25, 2026. The descriptions of third-party policies above reflect their published information at that time and may change.