Cooking up the perfect prompt

A recipe for lenders using GenAI

Savvy lenders are using tools like Lulu as a key ingredient in gaining a competitive AI edge. Lulu—the first lending intelligence companion—is a generative AI tool that uses natural language processing (NLP) to lower the technical barrier to advanced data analysis, industry benchmarking, and proactive strategic insights. Just by entering a few prompts into Lulu, you get smarter, simpler, and swifter insights into your lending operations.

For those new to prompting with GenAI tools, think of it like preparing your favorite dish; following the right recipe will give you the best results. And with LuLu as your sous chef, you can easily cook up some 5-star insights to support your business.

Much like cooking, AI prompting is a balance of art and science. While it can feel like an intimidating process, practicing helps you gain confidence and will get you the best results faster. Organizations across countless industries have growing appetites for AI and are recognizing that upskilling their teams should be a high priority. It is no different for the lending industry. A recent survey with +1,400 C-suite executives shows that only “1 – 10%” of employees are trained in GenAI tools despite the increasing interest and demand.

For lenders, using LuLu bridges the gap of turning raw data sets into fast and accurate lending intelligence. Reporting that used to take specific technical skills, countless hours of parsing through data, and even more time to discern valuable insights can now be done in a few minutes by asking a few questions. To get the most out of this technology, understanding the fundamentals of how to ask these questions is key. The recipe every lender should follow, especially with more complex analysis, is just four steps:

1. Goals – Clearly state what you want to accomplish, including the data and insights you need.
2. Formatting – Specify how you want the response formatted.
3. Caveats – Include details to flag or double-check
4. Context – Give some background or examples of what you’re looking for.

Putting all of these ingredients together ensures LuLu can deliver focused responses that help you surface the right data to guide you. To demonstrate, we’ll break down how a credit union could use LuLu to conduct an analysis of delinquencies amongst a peer group.

State your goals

Like with any quest for knowledge, you have an idea of what you are looking for, so it’s just as simple as stating what you want to know. Your goals should be specific and supplemented by data or insights you want to uncover. Starting an analysis on delinquency rates could look something like this:

Goal: I want to understand how our credit union’s delinquency rates for auto loans in Q3 – Q4 2024 compare to other similar-sized credit unions within the state of Virginia in that same time period.

Specify the desired formatting

With any report you want to share, it’s important to specify if the response needs to be formatted as bullets, graphs, a table with headers, etc. The good news is that the format can be changed after the fact with a simple follow-up prompt. 

Formatting: Format this as a table and include the name of the credit union, its asset size, its delinquency rate, and the percent difference between our delinquency rate.

Include caveats

Caveats could include details that LuLu needs to double-check to avoid common misinterpretations or understand the parameters it needs to follow. 

Caveats: Be careful not to reference “ABC Credit Union” and only include credit unions with asset sizes greater than $50 million.

Add context 

The amount of context can vary depending on the task. A rule of thumb here is to include information that you feel would be relevant if you were making this request to a human. 

Context: For context, our credit union experienced a significant increase in auto loan delinquencies in Q3-Q4 2024. We would like to understand if this was a common pattern with other credit unions in that time period. 

One of the best things about GenAI tools like LuLu is that, like cooking, the greatest flavors or ‘a-ha’ moments happen when you play and experiment. Stay curious, and you will push the boundaries of what you can discover while building confidence in your prompting. 

Conclusion

For lenders, Lulu will be the go-to resource for streamlining data analysis, surfacing meaningful insights, and other time-consuming but business-critical processes. In addition to effective prompt training, equipping teams with technology that understands the language and context of the lending industry will make a world of difference in the outputs.

Admit it—you want to try this prompt yourself. Sign up for a free 30-day trial of LuLu Pulse and give it a try!