AI is revolutionizing how fast companies are building products – and compliance is at risk of being left behind. Last week, a senior product manager Slacked me: “With how fast this environment is changing, being able to quickly test new [stuff] is proving to be super important. Almost feels like table stakes”
For tech product teams, continuous innovation is now table stakes. Solutions that were cutting-edge for a year may now last months or weeks. And it’s only getting faster. Successful companies in the tech space need to be nimble, flexible, and able to incorporate new technologies as they arrive.
This rapid innovation is not mindless, but a necessary reaction to the demands of the marketplace. Businesses are facing relentless external pressures caused by AI innovation. For example, a recent survey by Accenture found that 80 percent of cybersecurity executives at large global banks believe that generative AI is empowering attackers faster than banks can respond.
As the leader of my company’s compliance function, I think of these challenges differently. For compliance – my worry is figuring out how to move at the same speed as the rest of the business.
To succeed, it is clear that the compliance function needs to adapt. I don’t believe there is a fixed, prescriptive solution to navigate compliance in an AI world. The new reality is that any playbook is continuously subject to being rewritten.
That said, here are some compliance approaches I’ve found helpful to manage accelerating business timelines.
- Take advantage of technology: Compliance teams should be leveraging AI tools, whether it’s a large language model like Gemini, ChatGPT or Claude, or a more targeted solution, to buttress and amplify their capabilities.
- Move to Real-Time Risk Assessments: Compliance is shifting from periodic reviews to a continuous process—powered by dashboards and real-time alerts. We use Looker dashboards to support an array of real-time monitoring. Teams should consider how to embed more frequent model updates into their model risk management reviews (e.g. SR 11-7). It is also important to maintain a live inventory of the AI/ML models in use.
- Use frameworks as anchors: Regardless of the technology, companies need to adhere to core statutory and legal principles such as fair lending, protection against Unfair, Deceptive or Abusive Acts of Practices, privacy, and data security. The NIST AI Risk Management Framework, in particular, is useful in managing AI Risks.
- Stay close to your product and engineering teams: To advise product development, I need to know what product is building, so I sit in on weekly team meetings, join Slack channels, and participate in product development calls. It is critical to get compliance teams involved early in the conversation to avoid pitfalls, so we can infuse compliance considerations into new products.
- Leverage your industry groups: Financial services firms frequently participate in industry consortiums and working groups that collectively monitor AI developments and establish best practices. I have also leveraged outside groups like industry associations and professional groups (e.g., TechGC) to crowdsource best practices.
As a compliance officer in the age of AI, I find the pace of innovation both terrifying and exhilarating. But I’m convinced that by embracing real-time tools, embedding compliance in product design, and staying plugged into industry networks, compliance can become not just a safeguard—but a partner in innovation.
