AI Responsibility

AI is a tool in our work. It shortens research, drafting and repetitive project tasks so our time goes where experience matters: understanding how a business operates, making architectural decisions, and building something the team can run without us.

It does not decide anything. Every recommendation that reaches a client is the responsibility of a named person at CWT Studio.

• Drafting documentation, process write-ups and internal notes.
• Summarizing meeting notes into action items.
• Background research on platform behaviour and vendor documentation.
• Reviewing and generating code and configuration during a build.
• Editing our own written material for consistency with our editorial standard.

• Architectural decisions, scoping and sequencing. Those come from experience with the specific business.
• Assessment of a client's situation. Findings are written by a person who examined the systems.
• Any judgement about a person, including hiring, evaluation or eligibility. We do not run automated decision-making with legal or economic effect on individuals.
• Client testimonials, case study claims and metrics. These are verified at the source.

Nothing produced with AI assistance leaves our hands unreviewed. A person reads it, corrects it, and takes responsibility for it. If a deliverable is wrong, the accountability is ours and the remedy is in the Terms of Service. Pointing at a tool is not an available answer.

We do not submit client data to a tool that would train on it. Where we use AI tooling on material connected to an engagement, we use accounts and settings that exclude the content from model training, and we keep confidential material out of tools that cannot make that guarantee.

Personal information, credentials and identifiable client records stay out of prompts. Where context is needed, it is reduced to what the task requires, with identifying detail removed. Our handling of personal information generally is in the privacy notice, and controls on access are in the security overview.

Speed is not accuracy. AI produces confident text that can be wrong about how a platform behaves, so every configuration, automation and report is tested against the actual system and the documented requirement before it is delivered. Statistics and vendor behaviour cited in our writing are checked at the source. Where something cannot be verified, it is removed.

Articles and pages on this site are written and edited by a person, with AI used for drafting and editing support. Case studies describe engagements we delivered. Metrics are taken from client records or platform reporting, and anonymized where a client has not agreed to be named. We do not publish generated examples, generated quotes or generated results.

A client may restrict or prohibit the use of AI tooling on their engagement. Tell us and we work without it. Where a client's own policy governs AI use, that policy applies to our work for them. We disclose the categories of tooling in use when asked.

This area changes quickly. We review the tools we use, the data settings on each account, and this policy at least annually, and sooner when a vendor changes its terms. We adopt a tool where it improves delivery quality, and drop it where it adds risk or moves judgement away from a person. The effective date at the top of this page reflects the last review.

• Email: shannon@thecwtstudio.com
• Organization: CWT Studio
• Location: Ontario, Canada