If your team deploys AI that influences hiring, credit, housing, or insurance decisions in Colorado, you'll need a compliant notice framework by January 1, 2027. This template provides a foundation for meeting the disclosure and consumer rights requirements under SB 189, Colorado's Automated Decision-making Act.
Purpose of the Template
Colorado's ADM Act imposes three core obligations: notice of use, post-adverse-outcome disclosures, and consumer rights to access, correction, and human review. This template focuses on the notice and disclosure requirements, the customer-facing communications needed when your automated decision-making technology (ADMT) makes or materially influences a consequential decision.
Use this template when:
- Your system makes automated decisions about employment, housing, credit, education, healthcare, insurance, or legal services.
- The decision is consequential, affecting access to or cost of these services.
- You're the deployer, not just the developer.
This template doesn't cover developer documentation or the technical impact assessment requirements for those building ADMT for others.
Prerequisites
Before implementing this template, confirm:
Scope applicability: Ensure your system meets the ADM Act's definition of ADMT and produces consequential decisions. The Colorado Department of Justice's rulemaking process will clarify edge cases. If your system influences access to employment, credit, housing, insurance, education, healthcare, or legal services, assume you're covered.
CPA alignment: Review your Colorado Privacy Act (CPA) obligations. The ADM Act and CPA both address automated decision-making. Your notice framework should satisfy both where possible.
Human review capability: Have a process for meaningful human review of adverse decisions. The template includes language directing consumers to request review, but you need the operational capability to deliver it.
Data access infrastructure: Ensure you can provide consumers with the personal data used in their adverse decision within 45 days. This may require coordination between your ADMT system logs and your CPA data access workflow.
The Template
Standard ADMT Use Notice
Place this notice where consumers encounter your service (application forms, account dashboards, service agreements):
Notice of Automated Decision-Making
We use automated decision-making technology to [describe purpose: evaluate applications, determine eligibility, assess risk]. This technology analyzes [describe data categories: employment history, credit information, application responses] to make or assist in making decisions about [service type: employment, credit, insurance].
If our automated system produces a decision that adversely affects you, we will provide you with:
- A statement of the decision and its principal reasons
- Information about the data used and its source
- Your right to access and correct that data
- Your right to request human review of the decision
For questions about our use of automated decision-making, contact [designated contact method].
Post-Adverse-Outcome Disclosure
Send this within [specify timeframe based on your service type] after an adverse decision:
Automated Decision Notice
Our automated decision-making system has made a decision that affects your [application/account/service]:
Decision: [Specific outcome: application denied, rate increase, account restriction]
Principal Reasons: [List the primary factors that contributed to this decision. Be specific: "credit utilization ratio exceeded threshold," not "creditworthiness concerns"]
Data Used: The decision was based on [data categories: credit report data from Experian, employment verification from [source], application responses].
Your Rights:
You have the right to:
- Access the personal data used in this decision
- Correct any inaccurate data
- Request human review of this decision by a person with authority to reverse it
To exercise these rights, [provide specific instructions: submit request through account portal, email [address], call [number]].
We will respond to your request within 45 days.
Human Review Request Form
Provide this form or equivalent intake process:
Request for Human Review of Automated Decision
Your Information:
- Name: _______________
- Account/Application ID: _______________
- Decision Date: _______________
Request Type (check all that apply):
☐ Human review of automated decision
☐ Access to personal data used in decision
☐ Correction of inaccurate data
Explanation (describe why you believe the decision should be reviewed or what data you believe is inaccurate):
Supporting Documentation (attach any relevant documents):
By submitting this request, you are asking [Company] to conduct a human review of the automated decision identified above. A qualified reviewer with authority to reverse the decision will evaluate your request and respond within 45 days.
Customization
For your service type: Replace bracketed placeholders with specifics. If you're a lender, specify "credit application" and "debt-to-income ratio, payment history, credit utilization." If you're an employer, specify "employment screening" and "background check results, skills assessment scores, application responses."
For your data sources: List actual third-party data sources. "Credit report data from [bureau name]" is more useful than "credit information from consumer reporting agencies."
For your review process: Describe your actual human review mechanism. If you route requests to a loan officer with underwriting authority, say that. If you have a dedicated appeals team, name them.
For CPA harmonization: If you already provide automated decision-making notices under the CPA, merge the ADM Act requirements into that notice rather than creating duplicate communications. Proactive alignment reduces consumer confusion.
For materiality thresholds: The ADM Act applies when technology "materially influences" a consequential decision. If your system generates scores that a human always reviews before deciding, document that workflow. Your notice should reflect the actual decision-making process, if a human makes the final call using ADMT output, say "assists in making decisions" rather than "makes decisions."
Validation Steps
Before deploying this template:
Legal review: Have Colorado counsel confirm your interpretation of "consequential decision" and "materially influences" for your specific use case. The Department's final rules may clarify these terms.
CPA crosswalk: Map your ADM Act notices to your CPA automated decision-making disclosures. Identify overlaps and gaps. If you're already providing opt-out rights under the CPA, confirm whether those satisfy ADM Act human review requirements.
Operational readiness: Test your human review process. Can you deliver meaningful review within 45 days? Can your reviewer access the same data the ADMT used? Can they override the automated decision?
Data access workflow: Verify you can extract the specific personal data used in a given decision. If your ADMT system doesn't log input data at the decision level, fix that before January 1, 2027.
Consumer testing: Show the notices to a sample of your customer base. If they can't understand what data you used or how to request review, revise.
Monitor the Department's rulemaking process through late 2026, especially guidance on the ADMT definition, the "materially influences" standard, and CPA interoperability. Update this template as final rules emerge.
Your human review process is the operational commitment behind these notices. The template creates the disclosure framework, but compliance depends on your ability to deliver access, correction, and review when consumers ask for it.
