Consulting at marginal cost: what happens when your agent does the production

Marginal cost is the cost of serving one additional client once your infrastructure is built. For traditional consulting, marginal cost is high: each new client requires nearly the same amount of human time.

With an AI agent that encodes your methodology, the equation changes. The agent handles intake, qualification, first-draft production, and formatting. Your time goes to supervision, judgment, and approval. The production cost per deliverable drops because the agent does the repeatable work.

A consultant who previously delivered one audit per week can now supervise five, because the agent handles the production layer. The marginal cost of each additional audit is the supervision time, not the full production time. The expertise is the same. The economics are fundamentally different.

The shift is not about charging less. It is about changing what you sell. Instead of selling days, you sell supervised deliverables at a price that reflects the marginal cost of production plus the value of your expert judgment.

A security audit that takes a consultant two full days at 700 euros per day costs 1,400 euros. The same audit, with the agent handling the scanning, structuring, and first-draft analysis, takes four hours of supervision. At 150 euros per task, the client pays less, the consultant earns more per hour of actual work, and the throughput is higher.

This is not a race to the bottom. It is a structural change in how expertise scales. The consultant who encodes their methodology into an agent creates a production capacity that was previously impossible.

At scale, the consultant's role shifts from producer to supervisor. You define the methodology. The agent applies it. You review the output, catch what the agent misses, and apply judgment where the situation is ambiguous.

This is not a lesser role. It is a higher-leverage one. The decisions you make at the supervision layer, what to include, what to flag, what to reject, are the decisions the client is actually paying for. The production layer is necessary but commoditized by the agent.

Temet makes this concrete. The CLI audits your sessions, the kernel captures your methodology, and the relay delivers requests to your agent. Each component exists so that you can operate as a supervised production system rather than a manual service.

This model works for consultants whose expertise can be expressed as repeatable methodology. Technical audits, code reviews, strategic analyses, compliance checks, content strategy, financial modeling. Work where the framework is stable but the inputs vary.

It does not work for consulting that is primarily relational. Change management, executive coaching, organizational design. Work where the value is in the human interaction itself, not in a deliverable. If the client is paying for your presence more than your output, an agent cannot replace that.

The honest assessment: the marginal cost model amplifies consultants who have strong, repeatable methods. It does not help those whose value is primarily in showing up.