Jamilex Gotay
NACM National
Artificial intelligence (AI) has taken on a life of its own, integrating itself into almost every facet of our daily lives. Over the years, its use has only increased in both professional and personal settings. However, AI is only as useful as the information you provide it, and without the right context, it can generate false or inaccurate responses.
Why it matters: Credit professionals who learn how to best leverage AI are the ones who will stay ahead. By using effective prompting strategies, they can tailor it to deliver more accurate and actionable results.
How Credit Professionals Are Using AI
What started as an administrative tool to research customers and draft letters has evolved into a means of streamlining various credit functions. Neural networks can instantly organize, review, and generate large volumes of data, a feature that’s proved advantageous for credit professionals. “I present financial documents to Copilot—absent of customer-identifying information—and direct it to calculate specific ratios, summarize sections, pinpoint up or downtrends, and consider the credit request (by amount, terms, and industry),” said Krissy Aeschbacher, credit analyst at Maverik Country Stores, Inc. (Salt Lake City, UT).
Some credit professionals use intelligent systems for credit approvals or establishing credit limits for customers. “In Copilot, I’ll input the appropriate credit application fields along with the corresponding credit report of the customer to evaluate risk level,” said Haley Hagin Fimbres, CBA, CICP, credit manager at Superfeet Worldwide LLC (Ferndale, WA). “I use it to review contracts for unfavorable terms or missing information and to check if there is a mutual indemnification clause.”
However efficient generative AI can be, it’s not a perfect system. Its tendency to produce false or inaccurate information makes it a liability at times. Refining prompting strategies will help you receive the most reliable AI assistance possible.
Optimizing Your Prompts: Specify and Over-inform
Being specific and providing information in your prompts will help you get exactly what you want, when and how you want it. Providing details about your role, company, industry, and customer can provide sufficient context for more intuitive answers. “I’ve told AI to write me a professional, but semi-firm, collection letter for a customer that is X number of days past due that I want to retain in the B2B world,” said Sheila Roames, CCE, western regional credit manager at Ergon Asphalt & Emulsions (Chandler, AZ). “I’d also include that I work in the asphalt industry and ask for a reference to prior reminders and request payment within X number of days. I’ve found that providing this type of context, as well as my desired outcome, helps generate relevant content.”
Telling your AI platform what you don’t want can help enhance results and save time from making multiple revisions. For example, you can end your prompt by saying, “Avoid this” or “Do not include this.” You can even go a step further and preface your prompts with “For future reference,” “From now on,” or “Moving forward,” to guide future responses.
Use Examples
Using an example or a template can give the AI a reference point to help improve results. For example, if you’d like it to emulate a credit application, just say, “Follow the structure of this example,” “Draw from this,” or “Using the X attached.” However, it’s very important not to infringe on copyright. Don’t upload copyrighted works and ask the AI to rewrite/reproduce them or create a substantially similar work.
Provide and Ask for Feedback
Providing and asking for feedback will enhance the software’s findings and help you formulate better prompts. For example, you can ask for advice on how to get to a desired output, including what information is most useful and what prompts would be most effective. For instance, you can say, “Critique my prompt from the perspective of producing high-quality credit risk outcomes,” followed by, “Be specific about what’s vague or missing.”
Your guidance will enable AI to come up with better solutions that align with your needs. “It helps to treat it like a conversation,” said Aeschbacher. “I’ll write something like, ‘This is helpful, but it’s not quite right. Please remove X, adjust the tone to Y, and expand on Z.’ That kind of iteration always gets the document closer to our company’s voice.”
Treat it Like a Human
One way to generate more compelling outputs is to prompt AI as though it were a person, which in this case, would be a credit professional. Instead of asking, “What are the risks of extending credit to a new customer?” try, “If you were a seasoned credit manager with 20 years of experience in the manufacturing industry, what red flags would you look for before extending credit to a new customer?” By assigning a role, you give the AI a framework for tone, perspective, and expertise to produce responses that feel more grounded, specific, and human.
The bottom line: Prompting AI takes practice, but the investment is worth it. Credit professionals who treat AI as a collaborative tool, rather than a simple search engine, will find it capable of handling everything from drafting collection letters to analyzing financial statements.
For more information about Subchapter V, keep an eye out for the July/August issue of Business Credit magazine and a special session at NACM Connect’s Fall Conferences, Reconnect Live.
