Don’t Let Data Paranoia Hamper Your Bank’s Use of GenAI
By Ben Udell, SVP, Product Marketing and Innovation at Marquis
Simple Subscribe
Subscribe Now!
Executive Summary
- Reality check: You do not need private data or big tech investments to start finding value with GenAI in your bank or credit union.
- Nor should compliance fears or mishandling of sensitive data stop you from trying tools like ChatGPT or Microsoft Copilot. Your most valuable data — the kind that drives better decisions, saves time, and supports strategic planning — is right at hand and completely safe to use.
- Focus on readily available internal business data that is already anonymized or doesn’t reference specific individuals, or use publicly published data, to help with your decision making.
You don’t need a data science degree, or access to sensitive customer or member information, to start using GenAI in your bank or credit union. In fact, some of the most powerful use cases start with the operational and business data you already have: marketing reports, branch performance metrics, financial summaries, or product utilization trends.
GenAI can help you make faster decisions, uncover new insights, and build a more data-informed culture. But it starts with using these tools responsibly and intentionally.
Anna Khalzova, co-founder of Jupid, an AI platform that learns business context from transactions, counterparties, and conversations with business owners to make financial management simpler, works with data and AI every day. “I use GenAI not just to get answers, but to learn faster, work more efficiently, and move our business forward. As a founder, my time is limited, and AI is a force multiplier that everyone can use to work better. I have an always-on collaborator that helps me think through problems, test ideas and make better decisions in real time.”
You too can build a strong foundation for using GenAI at your financial institution (FI), no customer or member PII required. We’ll walk through how to get started, how to explore and learn even if you’re not data fluent, and how to apply GenAI in practical ways across marketing, retail, analytics, and more. (You can copy, paste, and modify these prompts to help you power your productivity and creativity.)
Want more insights like these? Check out MX’s content hub: Data in Action
Building a Responsible Foundation for Generative AI + Data in Banking
First, you don’t need customer, member, or proprietary data to get value from GenAI. You also don’t need to make large investments that are time- or resource-intensive.
Let’s say that again, because it’s the part that unlocks so much opportunity: You do not need private data or big investments to start finding value with GenAI in your bank or credit union.
Anna noted, “Working with small business data at Jupid has shown us that the most valuable insights often come from patterns, not profiles. You can learn a lot from how money moves, how processes flow or how customers interact. None of which requires exposing personal information. The key is using AI to understand context, not identity.”
For many banks and credit unions, the fear of violating compliance rules or mishandling personally identifiable information (PII) has been a major barrier to trying tools like ChatGPT or Microsoft Copilot. But the good news is that some of your most valuable data, the kind that drives better decisions, saves time, and supports strategic planning, is completely safe to use.
What Types of Data Can I Get Started With?
Focus on internal business data that is already anonymized or doesn’t reference specific individuals or use publicly published data that helps with your decision making. Examples could include:
Branch or regional performance reports such as loan volume, deposits, transaction counts, your Call Report, research documents, or government or trade association publications.
Marketing campaign performance such as open rates, clickthroughs, cost per acquisition, and ROI.
Operational metrics that help you manage your business, like call center response times, turnaround on loan processing, satisfaction scores, or feedback metrics.
Product utilization trends such as how often services are being used across different channels and engagement rates.
This type of data is already being used for business decisions, GenAI just helps you move faster and extract insights more easily. “For many banks and credit unions, they don’t need another dashboard. They need to understand the ‘why’ behind the numbers.” Anna also believes, “That’s what makes generative AI so powerful. It teaches you to ask better questions, even if you don’t speak ‘data’ fluently.”
Even though many GenAI tools now include appropriate protections (especially enterprise versions), it’s critical to understand your guardrails and use the right tool. Do not include PII unless explicitly allowed by your FI. Stick to aggregated, anonymized, or operational data that supports decision-making at the team or organizational level.
Read more from Ben Udell:
- We’re Speeding Past the Tipping Point of Gen AI in Banking
- The Silent Competition Is Winning: Are You Even in the Game?
- Four AI Action Plans to Supercharge Your Marketing Team’s Impact
How to Get Your Data into GenAI Tools the Right Way
The mechanics of working with data in GenAI platforms like ChatGPT or Microsoft Copilot are simpler than they seem, but how you format and present the data makes a big difference in how well your GenAI tool can help you.
When providing data to a GenAI tool, you typically have several options: uploading a spreadsheet, copy-pasting data directly into the chat interface, or attaching files. Uploading or attaching files allows for larger datasets and preserves formatting, making it easier for AI to process complex tables and multiple sheets. Copy-pasting is useful for quick analyses or sharing smaller datasets.
Here are a few foundational tips to get started:
1. Keep It Clean and Simple
AI works best with well-labeled tables, aim for clear column headers and rows.
2. Use Descriptive Labels
Avoid vague column names like “Q1” or “Data1.” Instead, go for “Q1 Net Revenue,” “Branch ID,” or “Monthly Operating Cost.”
3. Start with a Specific Ask
Instead of saying “analyze this data,” be clear about what you’re hoping to learn. Try prompts like:
- “Compare deposit trends across branches.”
- “What patterns stand out in this call center performance data?”
- “Suggest three follow-up questions we should ask based on this sales data.”
- “Ask me questions to help you better understand the data.”
Use GenAI as a Partner in Thinking
The power of GenAI is amplified when it’s used as a thinking partner. You can ask follow-up questions, explore trends, and request visualizations or written summaries to help communicate your findings. You’re not just automating tasks, you’re accelerating insights with a powerful assistant.
You don’t need to be a data analyst to get smarter with data. And you don’t need to enroll in a course to start learning how to think more analytically. GenAI can guide you, explain what something means, and help you build confidence along the way. It’s like having a helpful data analyst sitting next to you who never gets tired of your questions.
Try prompts like:
- “Can you explain ‘cost of funds’ in simple terms?”
- “What’s the difference between a pivot table and a filter?”
- “Why isn’t this VLOOKUP formula working?”
These prompts will give you more specific insights than a standard search engine, but this is where the power of GenAI starts. Follow up prompts help you realize the full educational power to help you power your knowledge and decision making.
- “Explain standard deviation to me like I’m in sixth grade and help me understand how I can apply this to my checking account monthly performance and give me questions I can ask my marketing and retail team to better understand what’s driving results.”
- “My cost of funds is considered high. Walk me through the formula for cost of funds, describe what it means, and ask me questions to help me think more critically about how my marketing team can influence lower this rate while increasing deposits.”
- “I need to turn my data about my cost of acquisition from a pivot table into charts and graphs to provide insights for the board book. I’m not sure how I should best display the data and then how to use Excel to generate helpful and insightful charts. Help me!”
You become smarter, faster, with the help of your GenAI tutor. The larger prompts begin to engage GenAI as though it’s a person; your best output comes from ongoing engagement and not one-time questions. One of the most powerful learning habits AI encourages is curiosity. The more you ask, the more you’ll start seeing patterns and developing a more strategic mindset.
Try prompts like:
- “What questions should I be asking based on this data?”
- “If I wanted to improve performance next quarter, what trends should I look for?”
- “What are three things I could analyze in this report that would help with forecasting and impress my boss?”
These short prompts will create powerful insights. By treating the GenAI like a thinking partner instead of just a task-doer, you’re actively improving how you look at data and how you make decisions with it.
Practical GenAI Use Cases Across Your Bank or Credit Union
Once you’ve got the basics down, the next step is applying it to the real work you do every day. Whether you’re in marketing, retail, data analytics, or another department, GenAI can help you uncover insights, save time, and make better decisions faster.
Here are easy to use everyday use cases to help you start using GenAI, even if you’re not deep into data analytics. These prompts are easy to repurpose into other areas of your bank or credit union by modifying the request and continuing to engage with GenAI.
Anna shared where her biggest learnings come from. “Some of our biggest breakthroughs have started with a simple prompt, something like, ‘Why is this number going up?’ What’s amazing is how quickly that turns into a deeper understanding of patterns, assumptions and strategy. GenAI doesn’t just give you an answer, it teaches you how to think differently while elevating your knowledge.”
Marketing teams are sitting on a goldmine of data, from email campaigns and social media stats to website traffic and ad performance. GenAI can help you make sense of it all, faster than ever.
Prompt 1: I’m going to give you campaign performance data for three email campaigns. Each campaign includes open rates, click-through rates, and conversions. Summarize the overall performance. Highlight which campaign performed best and why.
Prompt 2: Below is a table with results from four digital marketing campaigns. For each campaign, I’ve included spend, impressions, clicks, and conversions. Please calculate cost per acquisition and click-through rate for each one. Then color code and rank the campaigns from most to least effective based on performance. Identify which channel appears most efficient and provide reasoning. Suggest one way to improve each underperforming campaign. Also, include any anomalies or trends you notice in the data. Finally, summarize this information in bullet points for a leadership presentation.
Prompt 3: Act as a marketing analyst. I’m going to share a spreadsheet containing results from five campaigns across email, social, and paid search. Each row contains campaign type, budget, impressions, clicks, conversions, bounce rate, and customer acquisition cost. First, calculate ROI for each campaign. Next, compare results across channels and highlight which type consistently delivers higher conversions. Then identify two insights that would help inform our next campaign. Suggest one hypothesis we could A/B test in our email strategy based on what you see. Note any inconsistencies or outliers in performance. Recommend how to allocate next month’s marketing budget based on your analysis. Write a short paragraph summarizing the data in plain language for an executive audience. Follow that with a bulleted list of five takeaways. Finally, propose two follow-up questions we should explore further using this data. Do this in steps so we can dive into each request fully before moving on.
Retail & Branch Operations: Performance Snapshots and Staffing Insights
If you manage teams across branches or retail locations, you likely already collect performance metrics, foot traffic, new accounts, wait times, staff productivity. GenAI can help you analyze this data and turn it into action.
Prompt 1: I’m going to share a report with monthly foot traffic, number of new accounts, and staff hours by branch. Identify which branches are underperforming and overperforming and give me your rationale on why. Suggest a few questions we should ask based on the data to understand the results when I talk to the manager.
Prompt 2: You’re a branch operations consultant. Below is data from five branches including foot traffic, number of accounts opened, total teller hours, and customer satisfaction scores. Analyze which branches are operating efficiently and which may need support. Highlight where performance is below average and identify potential causes. Then, suggest staffing or scheduling adjustments we could test. Also, recommend one way to increase account openings at the lowest-performing branch. Write a few questions we should ask to explore the data further. Finally, turn your insights into bullet points for a weekly leadership report.
Prompt 3: Act like a performance analyst reviewing branch-level operations. The table I’m about to share includes foot traffic, number of transactions, customer satisfaction score, and employee headcount for 10 branches. Start by ranking branches by overall operational efficiency. Identify any outliers, positive or negative, and explain what makes them stand out. Suggest three follow-up metrics we should gather to explain differences in performance. Recommend an internal best practice that could be shared from a high-performing branch to a lower-performing one. Write two questions we could ask branch managers to get more context. Create a summary paragraph for executive leadership explaining the current state of branch operations. Follow it with a chart recommendation (bar or line graph) to visualize either customer satisfaction or efficiency. Propose one initiative to test in the lowest-performing branches. Summarize three insights you gathered and how they could impact future staffing plans. End by generating two follow-up prompts that a manager could ask AI to continue the analysis on their own. Let’s tackle this request in steps so we’re aligned in your approach.
Analytics & Research: Exploratory Analysis Without the Overwhelm
Not every bank or credit union has a dedicated data team. With the right prompts, GenAI can help you do basic analysis, forecast trends, and even run light statistical modeling all using the data you already have.
Prompt 1: I’m reviewing year-over-year growth across several business units. The table includes quarterly revenue and expenses by business line. Summarize overall trends and highlight where we’re seeing consistent growth or decline. Make one suggestion for further analysis.
Prompt 2: You are a research assistant reviewing a dataset with three years of financial results, broken down by product line. Do a deep dive on each product to assess quarterly revenue, customer acquisition cost, churn rate, and average revenue per user. Identify which products are growing, which are flat, and which are declining. Note any seasonal trends. Flag any product that’s becoming less efficient to operate. Recommend two product areas that deserve further research. Suggest a high-level strategic takeaway we should consider. Write a plain language summary of your findings that could be shared in a team meeting.
Prompt 3: I’m sharing a data set that contains five years of operational metrics across our top product lines. These include revenue, acquisition cost, churn rate, customer satisfaction score, and net new customers per quarter. Start by identifying long-term trends in each of these categories. Are we improving, declining, or holding steady? Highlight one product with accelerating growth and one that may be losing traction. Provide two reasons why these patterns might be happening. Suggest a set of metrics to track moving forward to validate your hypotheses. Create a short summary paragraph suitable for a slide in an executive strategy session. Recommend a basic forecast model in Excel we could use to project growth for the next two quarters. Also suggest a way to visualize this data to show progress over time. Propose a few questions we should explore using customer survey data, if available. Generate a short list of follow-up actions we could take to investigate or capitalize on the trends. Lastly, write three prompts that would help someone explore the data further using GenAI without needing advanced analytics skills. Start by creating a game plan and outline how to tackle this request.
You Don’t Have to Be an Expert. You Just Have to Start.
GenAI isn’t reserved for data scientists, IT teams, or big tech budgets. It’s for marketing managers writing campaign recaps, retail leaders comparing branch performance, and analysts trying to forecast next quarter with more clarity. It’s for teams that want to spend less time crunching numbers and more time acting on them.
As Anna puts it, “The more I use GenAI, the more it pushes me to think critically, ask better questions and get to the ‘why’ faster. Everyone in banking can use AI to work fewer hours, produce better work and make a more meaningful impact on their customers. It’s not just automation; it’s collaboration with your information.”
The path to using GenAI responsibly starts with the data you already have and the questions you’re already asking. With the right prompts and a little curiosity, you can turn everyday data into real business intelligence. And you can do it without touching customer or member PII or needing advanced analytics training.
