Pub. 13 2024 Issue 5

• Improper resource allocation due to poor or misaligned business intelligence. • Mitigation strategies that are interrupted by automated responses. • Regulatory penalties associated with poor/improper usage of AI tools. Proceed carefully. The key is thorough preparation and governance. How Banks Should Prepare for AI Implementation Because of the extra scrutiny around AI and data privacy, banks should elevate the risk level to at least one before any implementations. For instance, evaluating a product like a chatbot without customer data would typically have a low-risk rating. However, because auditors and regulators scrutinize AI tools more closely, you will want to treat chatbot testing as a medium risk. If you’re interested in pursuing AI tools at your financial institutions, let’s walk through what you should do step by step: • Step 1. Identify Key AI Opportunities Ask your IT committee to identify any areas that could benefit from automation. For instance, HR may need AI help sorting job applicants, while customer service may want a powerful AI assistance added to the customer portal. Be proactive by asking for your organization’s AI wish lists now. • Step 2. Create or Update Your Policy for Acceptable Use of AI for Your Bank If your bank hasn’t developed an acceptable use policy for AI, now is the time. Ensure the policy includes oversight from your AI board and establish guidelines for future approvals. Update the policy whenever new AI tools are added. • Step 3. Request the IT Committee Propose a Business Case Ask your group to put their recommendations into writing for senior management, creating a clear case including use cases, budget and likely productivity gains for the proposed technology. • Step 4. Hand Over Your Findings to Senior Management to Evaluate and Verify the Case for AI While this adds an extra layer of approval, it’s also an excellent opportunity to help educate senior bank leadership on how AI impacts the bank. • Step 5. Once You’ve Identified What You Want to Do with AI, Determine How You Want To Do It With your business case in hand, evaluate available AI solutions. Once you’ve selected a vendor, perform thorough vendor due diligence. Assess what data will be gathered, who has access to that data, what other ways the data will be used and where the data will reside. Finally, conduct a risk assessment and have it approved by senior management. • Step 6. If Your Tool Passes the Test, It’s Time to Put Together a Deployment Plan Once you’ve completed due diligence and testing, it’s time to deploy the tool. Involve key team members from Compliance (ISO), IT and senior management to verify alignment and expectations. Together, build a thorough testing and validation period to ensure the tool works correctly. Then, deploy the new capability widely. • Step 7. Create a Regular, Repeatable Validation Protocol AI tools are evolving rapidly, and vendors may frequently update their data usage and privacy policies. Regularly review these policies based on your vendor risk rating. Continuously assess the effectiveness of the tool and ensure it remains compliant with data privacy regulations. Put in the Work Now and Reap the Rewards of AI for Banks Though implementing AI may seem like a heavy lift at first, the long-term benefits are undeniable. AI can significantly enhance efficiency, customization and business intelligence at your bank. At Integris, we’ve helped hundreds of banks build IT infrastructure designed to grow with them. Our cybersecurity experts are on hand to help you with audits, policy development and procedures on a consulting basis. Either way, we’d love to talk. Contact us at integrisit.com. Though implementing AI may seem like a heavy lift at first, the long-term benefits are undeniable. 34

RkJQdWJsaXNoZXIy ODQxMjUw