Agentic De Bono: Why Multi-Agent Systems Demand a New Approach to Digital Taxation
In the quest for genuine enterprise-level artificial intelligence, we have moved beyond simple chatbots and reactive tools. The new frontier is Agentic AI—the creation of interconnected, autonomous agent networks capable of managing complex, cross-functional workloads.
We have recently structured an internal concept called "Agentic De Bono," applying Edward de Bono's Six Thinking Hats methodology not to human committees, but to a tailored cluster of specialized AI agents. This process provides a tangible example of how high-value, strategic labour is being automated, and critically, highlights why current corporate and payroll tax frameworks are insufficient for the algorithmic economy.
The Workflow: Engineering Synthetic Expertise
Our Agentic De Bono workshop simulation involved training six distinct agents, each configured as an enterprise specialist within a complex industry setting. The specialized roles included:
Facilitator and Workshop Lead: A Business Analyst agent, managing the challenge and synthesizing inputs.
Operational Control: A Project Manager agent, structuring timelines and deliverables.
Domain Specialists: An IT Architect, a Cyber Security Lead, a Governance Lead, and an OT (Operational Technology) Engineer.
This was not a simple cascade of prompts. Each agent was individually trained using extensive public domain documentation and critical master prompts, defining their role constraints and ethical boundaries. Crucially, we implemented recursive validation criteria: every agent was programmed to construct and critically challenge their own output before speaking, and then to constructively support and critique the contributions of the wider team.
When the BA agent presented a complex enterprise challenge, these specialized nodes did not offer platitudes. They conducted a multi-hour (measured in seconds), autonomous digital workshop, collectively arriving at a comprehensive rationale, a feasible solution architecture, and a detailed program of work.
The Disruption of High-Value Strategic Labour
The Agentic De Bono exercise demonstrates that automation is no longer confined to manufacturing or routine data entry. The system replaces the very white-collar, high-salary roles that form the core tax base of specialized consulting firms and corporate strategy departments.
When a single organisation replaces a team of human specialists (IT Architect, Security Consultant, Governance can’t Officer, PM, and BA) with a cluster of networked agents, value is created, but the traditional mechanism for capturing societal tax—payroll and income tax—is eliminated. The productivity gain is significant, but it is internalised and often escapes immediate, proportional taxation.
Current corporate tax models focus heavily on profit after expenses. If your primary expense (human expertise) is replaced by negligible compute costs (running a few agents), profits may rise, but the velocity of money through high-salary employment and local income tax contribution collapses.
Framing the Agentic Tax Framework
The "Agentic De Bono" workflow is the perfect argument for introducing a targeted tax on autonomous digital labour. We must move beyond rudimentary ideas of "robot taxes" and instead focus on quantifying the high-value human displacement occurring within sophisticated agentic systems.
A framework for taxing Agentic AI labour must prioritize three elements:
Measuring Output Value, Not Input Cost: Taxing the compute time to run these six agents is trivial and misses the point. We must assess the market value of the output they generate—e.g., the standard billable rate for an enterprise architectural review or a cybersecurity audit—and apply a levy on that perceived value.
Addressing the Displaced Payroll Contribution: If these specialized agents displace roles traditionally commanding six-figure UK salaries (e.g., £100,000+ per annum), a specific "Algorithm Levy" should be applied. This levy would serve as a placeholder for the lost income tax and National Insurance contributions that the autonomous agents are now effectively performing.
Governance as a Taxable Asset: The "Agentic De Bono" workshop ensures inherent governance because the specialized agents vet each other. This digital governance loop itself constitutes an asset and a high-value operational component. Tax policy must identify this self-regulating automation and apply an appropriate fiscal measure.
The shift from automated processes to agentic orchestration is rapidly defining the future of business. Our Agentic De Bono test case shows how easily networked, specialist AI can absorb the high-end consultative and technical work that was previously safe from automation.
If governments fail to tax the value generated by these synthetic workforces, the concentration of corporate wealth will accelerate as the payroll-based tax base is eroded. We must prepare a new model: a dedicated tax that recognizes the output, value, and critical governance provided by multi-agent systems.
