Faster Footnotes with AI-Powered Decision Trees

At Tellen, our vision is to empower auditors with GenAI to streamline their workflows. We analyze auditor workflows and figure out how to build data pipelines to efficiently extract, transform, and deliver data so auditors can complete their procedures more accurately in far less time. We are currently building a new module for FootnoteAI, an AI-powered decision tree, to quickly draft footnote disclosures that are specific and accurate to each entity, even if such disclosures were never previously created for that organization.

By using this AI-powered decision tree workflow tool, auditors will follow a sequence of steps that require answering questions and uploading documents that are processed by AI models. This orchestrated workflow ultimately results and the AI-powered generation of a footnote disclosure where the auditor remains “in-the-loop” providing oversight and guidance throughout the process.
‍

Our decision trees will first discern the necessary information about the entity from the user. After the profile characteristics (industry, reporting framework, etc) are established, the user will be prompted to input further information and supporting evidence/documentation such that the AI model can generate a workable draft of a footnote.
‍

We view our decision tree tool as the logical progression of the collaborative work we have been doing with our design partners to build our FootNoteAI application. We started by building our FootnoteDB tool, which assists in searching for comparable disclosures from publicly available 10-Ks or from your firm’s proprietary collection of financial statement disclosures. Next, we will be rolling out our clerical accuracy tool that checks for internal consistency throughout the financial statements, including throughout the footnotes. Now we are designing our decision tree tool. We have a few more tools that will be added to the FootNoteAI suite before that application is fully built.
‍

We also anticipate going deeper into this decision tree tool as it matures. We will be able to build autonomous AI agents into the workflow that specialize in specific repeatable tasks. For instance, we will be able to add an AI agent that automatically produces lease amortization tables whenever a new lease contract is uploaded. We may create another AI agent that continuously checks the FASB website for any regulatory updates before generating a footnote. Ultimately, this AI-powered decision tree tool will go from very basic background information and supporting documents to fully prepared workpapers and disclosures in minutes. We believe that the economics of firms that utilize our AI-powered tools will be fundamentally altered. We are excited to bring developments like these to our existing design partners and look forward to bringing more firms on board as they choose to AI-enable their businesses.

‍