AI Limitations

AI systems sometimes produce incorrect, incomplete, or outdated answers, and they do it with confidence. Treat AI output as a helpful first draft, not as fact. For anything important, verify before you rely on it.

Where a client-facing system handles sensitive conversations or decisions requiring judgment, the written scope defines the human review, escalation paths, and system boundaries included in that implementation. Not every implementation includes real-time human review. We recommend reviewing AI-generated communications periodically and reporting anything that reads wrong.

Where a system we build interacts directly with people, it is designed to identify itself as an automated system when asked, and where practical at the start of the interaction. We do not build systems intended to pass as a specific human being, and we do not build systems designed to manipulate people into harming themselves or others, to discriminate unlawfully, or to deceive. If your project needs something close to one of those lines, we will say so and decline or reshape it.

The optional AI guide on this website is automated software. It is not Daniel, it cannot book meetings by itself, it cannot send email by itself, and it cannot perform the paid audit. It may refuse inputs that look like passwords, payment details, or other sensitive credentials. That filter is imperfect. If you choose to send a summary, you review the exact text and press send.

We do not guarantee specific revenue, savings, booking rates, or other business results. Any ROI figures we provide are estimates and ranges based on your inputs and conservative assumptions. Actual results depend on your business, your market, and your follow-through.

AI outputs are not legal, medical, financial, or other professional advice and do not replace licensed professionals. Where a project may encounter regulated or professional-advice topics, the written scope identifies any disclaimers, limitations, and human-escalation controls included in that implementation.

Where an implementation sends messages (for example, SMS or email follow-up in a system we built for you), the written scope identifies the consent, opt-out, quiet-hour, identification, and human-escalation controls included in that system. Not every solution includes messaging. Where it does, the defaults are:

• Messages are sent based on consent captured through your intake and customer interactions, as defined in your scope.
• Messages are informational, not promotional. Mixing promotional content into transactional messages can change the consent the law requires, so we keep them separate by default.
• Quiet hours and send windows are configured as specified in your scope.
• Opt-out handling is configured as specified in your scope, and opted-out contacts are not messaged again.
• You are responsible for using the system lawfully, including the calling, texting, and email rules that apply to your business and your customers, such as the Telephone Consumer Protection Act and the CAN-SPAM Act. We configure the compliance features agreed in the scope. You own how your business uses them, including keeping the consent records your intake collects.

Our systems run on third-party platforms and AI models (for example, OpenAI and similar providers), plus project-specific tools named in your scope. Those vendors can change features, pricing, behavior, or availability. During the active engagement and included stabilization period, dbar Labs will communicate material vendor changes it becomes aware of that directly affect the delivered scope. Work required because of vendor changes, pricing changes, outages, or API changes may require additional scope. Vendor changes and outages are not defects in our work; fixes that require new scope are quoted separately.

Your business data is used to deliver your service, as described in the Privacy Policy. We do not sell it, and we do not use your confidential information to benefit other clients or to train general models for other clients. Where a project sends your data to an AI provider, that provider's own data terms also apply, and your scope identifies the providers involved.

Deliverables may include material drafted or generated with AI tools, reviewed and configured by us for your use. Such material may not be protectable by copyright in some jurisdictions. Your rights in deliverables are described in the Service Terms. You are responsible for reviewing AI-assisted outputs before relying on them for consequential decisions.

If any of this is unclear, ask before you buy. daniel@dbarlabs.com. Plain questions get plain answers.