Prompting for Performance Part III: Applying Prompting Best Practices to Investment Research and Due Diligence
Check out our previous posts here: Part I and Part II
In this last post of our 3-part blog series on LLM prompting for financial research and due diligence, we provide more concrete examples within the context of a real investment workflow. Brightwave users can take advantage of pre-built templates, reports and blueprints to create the outputs that are most useful within their specific workflows. Within all of these use cases, prompting is the glue that holds it all together: ensuring that the relevant, accurate information is presented in the relevant format (be it an email, memo, or slide deck).
Document Review and Summarization
Due diligence often requires reading through thousands of pages of lengthy documents, understanding and absorbing information from both proprietary business documents (supplier contracts, customer data, etc.) and public filings (earnings reports, press releases, investor slides). By asking an AI tool to “summarize the key points of this document, highlighting any financial irregularities or risks,” analysts can extract the information most pertinent to their thread of research. The prompt should direct the AI to focus on specific elements (e.g. “summarize the contract focusing on termination clauses and liabilities”). This ensures important details aren’t lost in a generic summary.
Extracting Insights and Red Flags
A well-prompted AI can rapidly sift through sell-side research, press releases, or even contract terms to identify red flags about a company or management team. For example, “List any legal or regulatory issues involving Company X in the past 5 years and describe their outcomes” is a clear prompt that provides structured parameters for answer extraction. The prompt explicitly asks for issues and their outcomes, guiding the AI to provide a concise report of only pertinent information. According to AlixPartners, combining AI with expert domain expertise in this way “accelerates analysis, unlocks broader and deeper insights, and de-risks” the due diligence process.
Financial Analysis and Comparison
Analysts can leverage AI to perform preliminary financial analysis by prompting it with data. For instance: “Compare Company X’s last three years of EBITDA margin and revenue growth to its top three competitors. What does this say about its competitive position?” A prompt like this instructs the AI on exactly what analysis to do (comparison with competitors) and what insight to draw (competitive position). With a well-constructed prompt, the AI might quickly crunch through the figures (if provided or available in its context) and deliver a benchmark analysis that an analyst can then verify and refine. This frees up time to focus on higher-level interpretation and strategy.
Despite these advantages, human expertise remains irreplaceable. Prompting AI for due diligence is most effective as a collaborative exercise between the analyst and the platform or tool. By honing the skill of prompt engineering, investment analysts and deal teams can dramatically enhance the efficiency and depth of their research. From quickly synthesizing data to spotlighting issues that warrant closer scrutiny, a well-prompted AI acts as a force multiplier for human expertise. The key is to remain in control of the conversation: the analyst sets the direction with thoughtful prompts, and the AI follows. Those who learn to communicate their analytical goals clearly to AI will find that it responds with more insightful, relevant assistance – ultimately leading to smarter investments and more thorough due diligence. In the evolving landscape of financial services, mastering the art of prompting is becoming as important as mastering the financial models themselves. It’s a new competency for the modern era of finance, turning cutting-edge AI tools into a trusted research partner when guided by knowledgeable domain experts.
