Colorado AI Compliance Guide for Financial Services Firms

What is the Automated Decision-Making Technology Act?

Colorado's new rules implement the Automated Decision-Making Technology in Consequential Decisions Act and the Conversational Artificial Intelligence Services Act. The Colorado Department of Law released an interim draft of these rules on October 6, 2026.

The Colorado Attorney General must finalize these rules by January 1, 2027, which is the same date both acts take effect.

What are the key updates in the latest Colorado AI interim draft?

The Department of Law filed an initial draft on August 11, 2026, and later introduced significant, substantive changes to the interim version after receiving more than sixty comments. You can participate in the final rulemaking process by submitting comments before the deadline on October 26, 2026.

The formal hearing takes place on that day, and the department accepts comments until 11:59 p.m. MST, or until the end of the final day if the hearing continues. The department noted in its cover letter that it remains open to considering alternatives to its current definition of a financial or lending service.

How does the 'Materially Influences' rule work?

The interim draft includes Rule 6, which establishes a framework to determine when AI outputs influence a consequential decision. An output is presumed to influence a decision if it constrains an option set, sets a threshold, or produces a rank, score, classification, recommendation, prediction, or other inference that pertains to you.

This presumption also applies if the output is reviewed by a decision-maker or used to screen information and is consistent with the final outcome. You can rebut this presumption by demonstrating that the output was a de minimis factor. This applies when other factors independent of the AI output played a significantly larger role in the final decision.

How do new Colorado AI regulations impact financial services?

Research by the Cambridge Centre for Alternative Finance and partners shows that 81% of financial services firms have adopted some form of AI. While 52% of industry respondents use agentic AI to take actions and make decisions on their behalf, only 14% of firms report that AI is transformational to their organizational strategy.

Fintechs show higher adoption rates, with 47% at advanced stages and 57% using agentic AI, compared to 30% and 45% for incumbents. Despite this, 48% of regulatory authorities describe themselves as still exploring AI or not engaged with it.

Firms currently operate in a fragmented environment with no single federal law governing AI in US financial services. Instead, you must answer to a stack of overlapping rules that shifts based on your jurisdiction, sector, and specific use case.

What are the risks and enforcement expectations?

The Attorney General’s policy is to focus on egregious and intentional violations and to exercise appropriate discretion while entities work toward compliance. Although there is no formal enforcement grace period, the law requires the Attorney General to provide notice and an opportunity to cure alleged violations before taking action.

This requirement for an opportunity to cure will sunset on January 1, 2030.

How can AI improve financial operations?

Agentic AI can significantly improve efficiency, as seen when it reduced the time required for anti-money laundering investigations by more than half. This saved over two hours of labor per case and more than 4 million Canadian dollars annually for one financial intelligence unit.

Despite these gains, 44% of financial firms still rely primarily on manual processes to combat fraud. For accounts payable, best-in-class teams process invoices in about 3 days, while other firms take slightly more than 17 days.