Good morning. Writing effective prompts for generative AI models is becoming a fundamental skill in finance that employers increasingly value.
In the past, finance and accounting resumes focused on CPA credentials, analytics, and technical experience, said Sundeep Reddy, EVP and chief accounting officer at Salesforce, during a panel session at Workday Rising in San Francisco on Wednesday. “Now, it’s shifting to: How have you used AI? What are your prompting skills? Have you written prompt scripts for agentic AI?” he said.
Salesforce employees are being trained not just to adopt AI agents, but also to write precise prompts. “You’re not coding—you’re writing plain English instructions,” Reddy said. “It takes trial and error to find the right level of detail for querying.”
Reddy, a former auditor, emphasized the importance of controls and validation: “Is this grounded? Is this auditable?” He recommended a “gut check” when reviewing AI outputs, and he encouraged iterative prompt refinement for accuracy and reliability.
A recent study, “As Generative Models Improve, People Adapt Their Prompts,” describes a large-scale experiment with 1,893 participants. The research found that only half of the performance gains achieved by switching to a more advanced AI model were due to the model itself; the other half came from users adapting their prompts to take advantage of its capabilities.
