Anthropic, the company behind AI assistant Claude, predicted that artificial intelligence could replace up to 50% of entry-level jobs in fields like finance, consulting and technology within the next five years.
As machines become more capable, it's becoming widely accepted that many of these roles will at least be augmented by AI. In my experience working with generative AI tools, chatbots like ChatGPT or Perplexity have been very helpful with toilsome tasks like summarizing documents, cracking Excel formulas, and occasionally teaching me how to cook a steak after work.
Valuation professionals have long relied on spreadsheets, manual data entry and judgment-heavy workflows to deliver fair value conclusions. Whether it's a valuation of business entities or financial instruments, the process often combines fragmented documents, customized spreadsheets, and many hours of manual work.
Generative AI models may start to change all that by streamlining repetitive tasks and enabling professionals to focus more on high-value analysis.
From grunt work to generative workflows
A significant portion of a junior analyst's time is spent on interpreting shareholder documents, updating the capital tables, and drafting valuation memos, which are all tasks that could be augmented by generative AI.
For instance, a key step in performing a 409A valuation is to understand the distribution waterfalls stated in the shareholder agreements. Locally hosted AI models on the laptop, like Ollama, can now ingest legal documents, extract relevant clauses, and output structured and actionable summaries. A locally hosted model also enhances data confidentiality, compared with cloud-hosted models like ChatGPT. Analysts can now focus on reviewing AI's output for accuracy rather than getting buried in the minutiae.
AI-generated models for complex security valuations
There are several benefits of building valuation models in Excel, including intuitive visuals and ease of internal review. While AI can generate complex Excel formulas in seconds, the current Excel Copilot is limited to mostly text-based assistance.
However, when valuing complex financial instruments such as restricted stock units or options with market and service conditions, modeling the underlying instruments in Python or other languages often produces cleaner and higher-performance models.
Generative AI can interpret the legal documents that define characteristics of the underlying instruments and create actionable modeling plans. For example, if valuing restricted shares that will vest based on whether the 10-day average stock price meets a certain dollar threshold, generative AI is fully capable of modeling these in Python and producing the relevant visuals to be used in the valuation report. A Monte Carlo simulation that might take several hours to run in Excel using Oracle Crystal Ball can often be executed in seconds using a Python model generated by AI.
Training generative AI on historical deliverables
Drafting valuation reports and audit memos often consumes a disproportionate amount of analyst time, given that these documents usually follow a specific structure, tone and compliance language. By training generative AI models on a firm's historical deliverables, it's now possible to automate a first draft of these documents that aligns with a firm's established standards.
For example, a generative AI trained on prior valuation reports or memos can learn preferred formatting, phrasing and disclosure language. By providing key inputs like company financials and financial model outputs, the AI model can generate a draft that resembles the firm's voice and formatting conventions. This doesn't eliminate the need for human review, but it significantly reduces the effort required to go from financial model outputs to client-ready deliverables.
What this means for the profession
Will business valuation work be fully automated by AI? Unlikely in the short term. Human judgment remains central in the profession. But the nature of the work is evolving.
Tomorrow's valuation specialists will need more than financial expertise; they'll need to understand prompt engineering, and the strengths and limitations of AI tools, and integrate them effectively into their workflow. The teams that embrace this shift early will not only be more efficient, but also be better positioned to meet client deadlines and deliver defensible work products.
As we move into this new era, one thing is clear: Spreadsheets are no longer the final frontier of valuation. AI isn't here to replace valuation specialists — it's here to take the tedious parts of the job off their plate.
