AI will not replace the valuation professional. But the firms that understand what it can — and cannot — do are already pulling ahead.
Artificial intelligence is reshaping the financial services landscape rapidly, and the valuation profession is no exception. What began as spreadsheet automation has evolved into something considerably more consequential: AI systems are capable of analysing large volumes of financial information, drafting reports, interrogating models and performing multi-step analytical tasks with limited human intervention.
Recent announcements from Anthropic, the company behind Claude, underscore how quickly enterprise AI is being tailored to knowledge workers in financial services. Most notable is the launch of Claude Cowork — an agentic AI system designed for professionals who work with documents, data, and complex workflows every day. Unlike a conventional AI tool that responds to a single question, Cowork accepts a defined outcome and works autonomously through the steps required to get there. For valuation professionals across South Africa, the question is no longer whether AI will affect the profession — it is how to engage with it on the right terms.
The answer lies in a concept worth taking seriously: judgment augmentation. Not replacement. Not automation for its own sake. But a deliberate shift in how professionals allocate their time — away from manual processing and toward the interpretive, commercial, and relational work that AI cannot replicate.
“The most significant shift is not that AI can perform calculations faster. It is that a capable model can now hold the entire analytical context of a valuation — financials, industry data, comparable transactions, regulatory considerations — in view simultaneously, and reason across it.”
How AI is Enhancing Valuation Workflows
AI tools are creating meaningful efficiencies across the valuation process. Industry research that once took hours can now be completed in minutes — market trends, competitor positioning, and sector commentary synthesised from large volumes of documentation. Financial statement analysis is similarly accelerated: AI can extract key trends, flag unusual movements, and populate valuation models from source documents with considerably less manual effort. For comparable company and transaction analysis, AI connected to financial databases can screen, rank, and calculate multiples at scale — shifting the professional’s role from data collection to critical evaluation. And in report drafting, AI can prepare first-draft narrative sections, summarise methodologies, and standardise language consistency — freeing practitioners to focus on the commercial judgments that clients value most.
The Rise of AI Agents — and What It Means for Valuation Workflows
The next significant evolution is the emergence of AI agents. Unlike conventional AI tools that respond to individual prompts, agents are designed to perform multi-step workflows with limited human intervention. Anthropic’s Claude has been developed with an explicit focus on safe, auditable, enterprise-grade agent behaviour — properties that matter considerably in a financial services context.
In a valuation setting, an agent tasked with preparing a preliminary DCF analysis might gather the relevant financial data, structure the analytical framework, identify comparable transactions, run sensitivity analyses, and produce a draft report — each step traceable, each assumption documented.
Crucially, Claude is designed to articulate its reasoning — making agent outputs auditable rather than opaque. In a profession where conclusions must be supportable and methodology defensible, this transparency is a prerequisite for responsible adoption. While agent-driven workflows remain in relatively early stages of adoption, the trajectory is clear. Firms that begin building the internal capability and governance frameworks now will be significantly better positioned as these tools evolve — and each new generation represents a step-change, not an incremental update.
THE SOUTH AFRICAN CONTEXT
South Africa’s valuation landscape presents distinctive challenges that generic AI tools are poorly equipped to handle, including thin and illiquid trading markets, dual-listed structures, B-BBEE ownership complexity, rand volatility, and a regulatory environment spanning SARS, the FSCA, and JSE disclosure requirements. Valuations prepared in connection with Section 42, 45, or 47 transactions under the Income Tax Act carry specific methodology and disclosure requirements that a tool calibrated for other markets may not recognise. AI configured with South African-specific knowledge — including the Companies Act, JSE Listings Requirements, applicable SARS guidance, and POPIA obligations — can be substantially more useful than an out-of-the-box model. The opportunity for locally informed AI adoption is real; so is the risk of relying on tools that were never built with this market in mind.
Why Human Judgment Remains Indispensable
Despite the genuine opportunities AI presents, valuations remain fundamentally reliant on professional judgment. Valuations require nuanced consideration of company-specific risks, management quality, strategic positioning, transaction context, and the behaviour of market participants — weighing contradictory signals, applying appropriate scepticism, and reaching a defensible conclusion in the face of uncertainty. These are not tasks that can be delegated to a model, however capable.
AI systems may assist with processing information and improving efficiency, but they do not replace the need for professional oversight. Valuation professionals remain responsible for ensuring that conclusions are reasonable, supportable, and aligned with applicable professional and ethical standards — the IVSC’s International Valuation Standards being the most directly relevant benchmark. AI is a tool in service of that responsibility, not a substitute for it.
RISK, GOVERNANCE & CONFIDENTIALITY
Client Confidentiality
Valuation engagements involve highly sensitive financial information. Firms must ensure AI tools are deployed within environments that meet appropriate data security standards before any client data is processed.
POPIA Compliance
South Africa’s Protection of Personal Information Act imposes obligations on the processing of personal information. Firms should assess their POPIA obligations and ensure appropriate data processing agreements are in place.
Hallucinations & Accuracy
AI models can produce plausible but incorrect outputs. All AI-generated analyses must be subject to professional review before any reliance is placed on it. AI reduces effort; it does not reduce the obligation of verification.
Documentation & Auditability
Where AI tools are used in the preparation of a valuation, firms should document the nature and extent of that use. As professional standards evolve, demonstrating appropriate oversight will become increasingly important.
Looking Ahead: The Firms That Will Lead
The valuation profession will not be replaced by AI. But it will be materially transformed by it — and that transformation is already underway. Over time, professionals are likely to spend less time on manual data gathering and administrative drafting, and more time on interpretation, strategic analysis, and client engagement.
The firms best positioned to benefit are not necessarily those with the largest technology budgets. They are the ones that approach AI adoption thoughtfully: investing in the right tools, building governance frameworks, and training their professionals to use AI critically.
In a profession built on the credibility of its conclusions, that combination of capability and judgment is not just a competitive advantage. It is the standard that clients, counterparties, and regulators will expect.
