Entrepreneur Built Practice Staffed by AI Agents
Alexis Kingsbury detailed his management experiment involving eleven distinct AI agents in a September 2026 podcast.
Updated on Sept. 24, 2026 in Artificial Intelligence
Live Poll
Do you trust keeping organizational knowledge in AI-connected systems you do not personally control?
Alexis Kingsbury conducted a management experiment in September 2026, building an accountancy practice entirely staffed by eleven AI agents. The project, discussed on the Accounting Tech Lab podcast, served as a test for delegation, workflow design, and human control over automated processes.
Why it matters
This experiment highlights the critical necessity for firm leaders to maintain control over core organizational knowledge and processes when integrating automated agents. It demonstrates that while AI models can handle complex tasks, they are prone to simple errors, necessitating human-in-the-loop oversight.
The firm utilized eleven AI agents each configured with unique personalities, roles, and responsibilities. The system integrated these models with internally controlled platforms to preserve organizational context.
The players
Alexis Kingsbury
He is an author and entrepreneur who wrote the book Accrual Intentions.
Accounting Tech Lab
This is a specialized podcast series that explores the integration of new technologies into the accounting profession.
The details
Kingsbury designed the firm workflow with mandatory stage gates to facilitate human review of AI outputs. He advocates for a strategic separation of probabilistic AI tasks from deterministic accounting processes to ensure data integrity.
Timeline
September 2026: Alexis Kingsbury discussed the experiment on a podcast episode.
The Tech Race
This project represents a shift toward LLM-based autonomous agent workflows that replace legacy human-centric task execution. It positions these agent-based systems as the next phase of corporate automation beyond simple digital tools.
Users can expect to see more workflows incorporating AI agent oversight as firms adopt stage-gate review processes for automated output. This shift necessitates that employees focus on supervising deterministic logic rather than performing repetitive manual calculations.
The takeaway
Management of AI agents requires the same rigor as human delegation, specifically through well-defined controls and stage gates. Implementing these tools is most effective when firms keep deterministic processes separate from probabilistic machine learning outputs.
Further reading
For more information on the evolving landscape of automated labor, visit the Artificial Intelligence section.
Source note: This article includes information reported by CPA Practice Advisor.
Live Poll
Do you trust keeping organizational knowledge in AI-connected systems you do not personally control?
