The Hidden Volatility of AI-Driven Revenue

The accounting industry remains tethered to the assumption that customers are human, ignoring the reality that AI agents lack loyalty, memory, and economic sentiment. When procurement or coding agents make purchasing decisions based on patterns learned during model training, they bypass traditional business relationships. A shift in a foundational model—such as OpenAI retraining GPT—can redirect spending toward a competitor instantly, turning revenue streams into a winner-take-all volatility trap.

Companies must begin segmenting their revenue into human-driven, agent-driven, and hybrid transactions to provide investors with a realistic view of concentration risk. While management teams may cite the technical difficulty of tracking these non-human users, the status quo misrepresents risk and leaves shareholders vulnerable to sudden, invisible shifts in demand. Investors currently have few options other than applying discounts to companies that cannot articulate their exposure to these automated cycles. As agentic AI becomes a standard component of the software value chain, the ability to distinguish between durable human relationships and ephemeral agent integrations will define the next generation of due diligence.