The Wisconsin Bankers Association Webinar on Generative AI was an engaging and insightful webinar filled with questions from bankers eager to understand how generative AI (AI) can be integrated into their organizations. It’s clear that many are looking for practical ways to start leveraging AI in their institutions. This blog aims to provide helpful and straightforward answers to the questions raised during the webinar, encouraging bankers to take their first steps into the world of AI. Let’s dive into the questions and explore how AI can benefit your bank.
To help all of the attendees learn more about AI, I created this video. The video shows you how I took the questions and worked with AI to generate a draft of this blog. You’ll see how I took screenshots of the questions, used AI to detail them out, organize them, and continue to work through a process to reach a draft I edited in Word. You’ll see in action how I applied concepts from the Webinar, along with some new tips and tricks.
I’d encourage you to pay attention to the actual prompts I used. While the blog is a touch long, it’s generated from very limited prompting. The effort is the critical thought process to give AI direction.
The rest of this blog is the final version after I’ve edited ChatGPT’s output. AI is never the last step!
General AI Use and Adoption
AI has the potential to transform banking. During the webinar, several questions were raised about how to begin integrating AI, especially in organizations where there may be resistance or hesitation. Start with a practical approach and real examples to help you and your team understand the benefits and applications of AI, and how to start leveraging it effectively.
What are some use cases I can share with my IT team for letting the Marketing Department use AI? (Alternatively, how I can help IT understand how Marketing can use it.) They are fully against it and not very open to anyone else’s opinion.
AI can be a powerful tool for IT teams, providing numerous benefits that can enhance efficiency and productivity. Here are some key use cases:
- Writing Code: AI can assist in writing and optimizing code. This can speed up development times and help identify potential issues early in the development process.
- Error Checking: AI tools can help in debugging code by identifying errors and suggesting fixes. Simply ask AI to interpret a code and provide guidance. AI is excellent at solving these issues.
- Translating Technical Topics: AI can help translate complex IT topics into understandable language for non-technical stakeholders. This can improve communication between IT and other departments, making it easier to gain support for IT initiatives.
The webinar demonstrated some clear marketing wins. However, it can be important to highlight how big the productivity gains are with content creation. Focus on the productivity improvement, and you could ask AI to give you examples relatable to IT!
Does ChatGPT learn from prompts provided previously in other chats and start to learn what I’m looking for? Or, are we starting from scratch with each new chat?
ChatGPT does not remember information between different chat sessions. Like your text messages on your phone, each chat is different. However, there are some new and ongoing developments that can allow this to happen. Plus, check out ChatGPT’s GPTs if you need memory and consistency in your work.
First experience with ChatGPT – so FDIC regulations are available and AI can access and we could request it to create training for, say, Fair Lending using that regulation? Or would we need to “upload” that regulation?
AI has done a great job of understanding publicly available FDIC documents. I’ve found generally it’s great and spot one, with the standard disclaimer (and very appropriate for this question) you should fact check your work. You can upload or input the specific regulations you want AI to use if you’d prefer to dial in its focus. Try both methods and review the results.
I see you’re using ChatGPT-4. Is that the preferred AI option for someone just starting out in using AI?
I understand ChatGPT and have worked with it a lot to appreciate strengths and weaknesses. There’s value to that, but certainly try other systems. I’m really enjoying Gemini Advanced right now. Copilot is nice, but ChatGPT is my BFF.
Longer term I’d project most users will use Copilot because it’s embedded in Microsoft products and we all use them already. ChatGPT, Gemini, and others will be more limited to use users or use cases.
I see you are using ChatGPT for all of the examples. Have you found this to be the best site to use? Also, does your bank have any security concerns with using AI software?
We do not put personally identifiable information (PII) into ChatGPT, although their Teams and Enterprise models have solutions that do not train on this data and keep it private. I’d research that and make a decision based on your comfort level.
You’ll have no problem getting value at $20 and not using it for PII.
Great topic and great information! Are you able to come to my bank to train or even talk over the phone for more guidance?
Reach out, I’m open to helping others depending on the timing and needs!
If we’re just getting started with AI, how broad would you suggest we promote this? In other words, should we have a few key players start playing with this regularly or broaden the scope?
Starting with a pilot program involving a few key players is often a good approach. This allows your team to gain experience with AI, identify any challenges, and refine your strategy before scaling up. Once the initial team is comfortable with the technology, you can gradually expand its use across the organization. This phased approach helps ensure a smooth transition and minimizes potential disruptions.
Security and Compliance
When integrating AI into banking operations, ensuring the security and compliance of sensitive information is paramount. During the webinar, many participants expressed concerns about protecting confidential data, avoiding copyright infringement, and verifying the accuracy of information. Here, we address these critical issues, offering strategies for safeguarding information, using tools to check for plagiarism, and leveraging AI’s familiarity with regulations. Remember, AI can be a valuable tool in maintaining compliance and accuracy, but it’s essential to use it wisely and always ask for citations when in doubt.
How should banks protect confidential information or other bank proprietary information when using these features?
To protect confidential information when using AI features, banks should follow these best practices:
- Avoid Using PII: Do not input PII into AI systems unless it is explicitly approved and necessary. This minimizes the risk of data breaches and misuse.
- Data Encryption: Ensure that any data shared with AI systems is encrypted during transmission and storage. Review their Trust site information for more info.
- Access Controls: You may start with limited access. But remember, there are a lot of AI solutions out there and some are embedded in other solutions. It’s probably impossible to fully restrict AI, which means your first step is guardrails and expectations.
Do we need to be concerned about copyright infringement issues and potential plagiarism when using AI? How do you suggest mitigating these risks?
Yes, there are concerns about copyright infringement and plagiarism when using AI-generated content. To mitigate these risks:
- Use Plagiarism Checkers: Tools like Grammarly and other plagiarism detection software can help verify that AI-generated content is original and does not infringe on existing works.
- Review and Edit: Always review and edit AI-generated content to ensure it adheres to your organization’s standards and does not contain any copied material.
- Cite Sources: When asking AI to generate content, always request it to cite sources. This helps trace the origin of the information and ensures that proper credit is given.
Treat AI work like any other work you’re completing.
How do you fact check data and ensure it’s from a reputable source?
This is easy, simply ask for citations. You can take the links, verify the info, and ask for additional sources if needed.
Practical AI Applications and Management
Generative AI can streamline many processes and provide valuable insights across various lines of business. During the webinar, participants were curious about how AI is being used to analyze customer data and how its implementation is managed within an organization. Below, we provide answers to these questions, explaining current practices and how AI can be effectively integrated into your operations.
Are you using AI to analyze customer data to help with things such as reducing attrition, deepening relationships, etc.?
Currently, we are not directly analyzing customer data with AI. Instead, we rely on our vendors who specialize in this area and deliver excellent results. They use advanced AI tools to provide insights that help reduce customer attrition, deepen relationships, and enhance overall customer satisfaction.
You’re sharing a lot of great examples of how to use ChatGPT to save time for various Lines of Business. Who is in charge of managing the use of this? Is there someone designated in HR who would have access to create training materials? Then someone else from Digital Banking who wants to review the 5 hours of FISERV videos? Or do they come to your Marketing department or Digital department to ask for this?
Our approach to AI implementation is currently decentralized, allowing various departments to leverage AI tools according to their needs. Generally, the use and support of AI runs through me, ensuring that everyone has the guidance and resources they need. For specific tasks such as creating training materials or reviewing extensive video content, team members from HR, Digital Banking, or Marketing can all be involved, depending on their requirements. They usually reach out to me for support and coordination to ensure effective use of AI across the organization.
AI Responses and Accuracy
Understanding how AI generates responses and ensuring their accuracy are critical concerns for many banks. During the webinar, participants asked about the consistency of AI-generated answers and the rationale behind AI decisions. Below, we address these questions, offering insights into how AI responses can be tailored and verified to prevent concerns about identical outputs across different banks.
If 15 banks put in the same question, will they all get the same answers? Or how different will the answers be? Same questions, same prompts.
If multiple banks input the same question with the same prompts, the AI is likely to generate similar responses. But you’ll see that if you repeat your prompts, you’ll get different versions, validating no one will have identical results. You’ll be applying your brand, specific requests, and providing personalized direction. Then you’ll edit. Here are three key points to keep in mind as you personalize the content.
- Customization: Tailor the prompts with specific details relevant to each bank’s unique context. Adding context-specific information can lead to more personalized responses.
- Iteration: Encourage iterative refinement of the AI’s responses. By providing feedback and making adjustments to the prompts, banks can achieve more varied and customized outputs.
- Human Review: Have a human review and edit the AI-generated content to ensure it meets the specific needs and standards of each bank. This step can add a unique touch to the final output.
![](https://cdn.shortpixel.ai/stsp/to_webp,q_lossy,ret_img/https://robotsarepeople.net/wp-content/uploads/2024/06/Robot-Images-3-300x300.png)
Combating Negative Perceptions
How do we combat the negative stigma surrounding the usage of AI in general? It’s often seen as plagiarism.
Combating the negative stigma surrounding AI usage involves focusing on responsible and ethical practices, which help build trust and demonstrate the benefits of AI. Here are some strategies to address these concerns:
- Promote Ethical Usage: Emphasize the importance of using AI responsibly and ethically. This includes ensuring that AI-generated content is original, properly cited, and used to enhance rather than replace human creativity and expertise. By showcasing examples of ethical AI usage, banks can help dispel fears of plagiarism and misuse.
- Transparency and Education: Educate employees and stakeholders about how AI works and its potential benefits. Transparency about the AI’s capabilities and limitations can help demystify the technology and reduce misconceptions. Providing training and resources can empower users to use AI tools confidently and responsibly.
- Highlight Success Stories: Share success stories and case studies where AI has been used effectively and ethically to improve operations, customer service, and decision-making processes. These examples can demonstrate the positive impact of AI and encourage its adoption.
- Advocacy and Leadership: Be proactive advocates for AI within your organization and industry. Leaders should champion the responsible use of AI and highlight its role in driving innovation and efficiency. By leading by example, banks can influence broader acceptance and adoption of AI technologies.
- Implement Plagiarism Checkers: Use tools like Grammarly and other plagiarism detection software to verify the originality of AI-generated content. This helps ensure that the content is unique and not infringing on existing works, addressing concerns about plagiarism directly.
By focusing on these strategies, banks can effectively combat the negative stigma surrounding AI and foster a culture of responsible and ethical AI usage. This not only supports strong AI adoption but also positions the organization as a leader in innovative and ethical practices.
Wrap Up
The Wisconsin Bankers Association Webinar on Generative AI provided a valuable opportunity to explore the many ways AI can be integrated into banks. By addressing concerns about security, compliance, practical applications, and negative perceptions, I hope to have provided a clearer understanding of how AI can be used responsibly and effectively.
If you have any further questions or need support in implementing AI within your organization, don’t hesitate to contact me. Together, we can harness the power of AI to drive innovation and improve efficiency in your bank. Let’s embrace the future of banking with confidence and responsibility.
Reach out for more information, guidance, or to discuss how we can tailor AI solutions to meet your specific needs.