How a PE Firm Automated Their Company Due Diligence Process with AI

A private equity firm approached us to automate a highly labor-intensive process with AI: their company due diligence process.

Due to the sensitivity of their data, they sought to automate this process using AI on an enterprise-grade platform. They lacked internal AI expertise and needed a solution quickly.

They partnered with us and deployed the internal tool in less than two weeks. It is now used by hundreds of analysts across various geographies.

Opportunity

Each fund analyst manages an Excel model containing a long list of investment opportunities (i.e., companies). This model contains initial information extracted from a public database with annual accounts (i.e., P&L, balance sheet, etc.).

To make inform investment decision, analysts enrich their model by outlining the business model, adding the company's URL, determining if the company is sustainable, and identifying the sustainability vertical (e.g., energy, agriculture, climate).

Prior to using AI, each model was sent to a third-party vendor whose job consisted of filling out additional fields, a process that took around 2 weeks and cost $2-5k each time they needed to update the file.

Once received the enriched model, they had to go company by company and check the information and determine the investment attractiveness. Just this part of the process took them 5-10 hours every week.

Solution

They broke down the process into several steps, setting up a different LLM with its own prompt for each step.

Use Case Chain of LLMs

The system retrieved information from the company's Sharepoint for additional context on how to perform the tasks: how to determine if a company was sustainable, the different sustainability categories, how to gauge investment attractiveness from internal documents and presentations, etc.

The system was exported as a batch processing interface, allowing analysts to upload an Excel file with a list of companies and receive an enriched Excel file with additional columns.

By clicking run, the AI system processes all the rows (i.e., all companies) in the Excel file.

Batch Processing Interface

Results

A process that previously took more than two weeks was reduced to less than five minutes. It is now fully internalized by the team, eliminating the need for external vendors for this critical part of the investment process.

By automating this process, the PE firm analyzes more companies, finds additional opportunities, and responds to market changes faster, increasing the fund's performance.

This initial AI use case, deployed in less than 1 month, sparked internal interest in automating more processes with AI.