AI Use Policy

Why We Use AI

AI has become part of the standard toolset for design, development, and marketing teams across our industry. We use it to speed up routine work such as early research, first drafts, code review, and bug triage. The time saved goes back into the work that depends on people: strategy, design decisions, and conversations with your team.

Our Principles

People make the decisions

AI supports our team on routine and repetitive tasks. Strategy, creative direction, and final calls on your project stay with people on our team and yours.

Client data stays protected

Our team does not enter client information into AI tools that use shared training data. When a task calls for sample data, we use anonymized data whenever possible.

AI work checked before it’s used

AI-assisted work is checked for accuracy and relevance before it becomes part of a deliverable.

Developers test the code

Code written with AI assistance goes through testing and validation by experienced developers before it ships.

Bias checked and corrected

We assess the AI tools we use for potential bias and correct course when we find it.

Approved org tools only

Our team works in COLAB-provided AI tools that have been reviewed for privacy and security.

Efficiency and Sustainability

AI runs on data centers that use significant energy and water. We pay attention to that environmental impact and promote responsible AI use across our team to reduce it.

Efficient use starts with shared habits. We maintain a library of prompts and skills that put our best practices into everyday work, and our team follows AI release notes to move to more efficient models and data sources as they become available. Experience matters here too. When our team can draw a conclusion from years of client work, we skip the lengthy AI session. AI has diminishing returns, and we make deliberate choices about when to use it and when expertise will get to the answer faster.

How AI Shows Up in Our Work

Here are some examples of how we use AI in our day-to-day work.

• Research and sales: We use AI for early research on an organization or industry before a first conversation. It is never used to segment or target people based on sensitive or discriminatory criteria.
• Marketing: AI helps analyze trends and draft content. We do not use it to make misleading claims or to play on people’s fears.
• Design: AI handles repetitive tasks such as resizing images and generating layout variations. Designers make the creative decisions.
• Development: AI helps improve code efficiency and detect security vulnerabilities. Developers review, test, and validate the result.
• Quality assurance: AI helps prioritize and categorize bugs. A QA professional reviews each issue before it is resolved or closed.
• Project delivery and product management: AI helps with scheduling and routine tasks like data entry. Feedback from your team gets a response from a person, and product decisions are made with your input.

What We Will Not Use AI For

Our team does not use AI to:

• Collect personal data without consent
• Create or reinforce bias and discrimination
• Produce misleading content, including deepfake images, video, or audio
• Conduct unauthorized surveillance or access personal or corporate data without permission
• Generate content that infringes on copyrights, trademarks, or patents
• Support cyberattacks, phishing, or any other illegal activity

Misuse of AI is handled by COLAB’s leadership team.

Your Organization’s AI Requirements

Some clients have their own AI policies, especially around patient, member, and customer data. We may be able to adjust our practices on a case-by-case basis, after our team reviews the impact on your project.

Governance

This policy is an addendum to COLAB’s Information Security Policy. Each approved AI tool has an assigned owner on our team, and AI-generated material is audited regularly. Our team receives training on AI tools and practices as they change.

Questions About AI on Your Project

To learn how this policy applies to your work, let’s talk.