Negotiating with AI Agents by 2028: Strategy & Governance

Key takeaways

• Agents will accelerate preparation and execution, not replace accountability.
• Define boundaries: what agents can propose vs what humans approve.
• Use evaluation and monitoring to prevent drift.
• Keep audit trails for decision legitimacy.

Where agents help

• Research and preparation
• Drafting proposals and counteroffers
• Scenario simulation

Where humans must stay in charge

• Final commitments and risk acceptance
• Exceptions and escalations
• Relationship and strategy

Governance blueprint

Combine policies (what’s allowed), controls (what’s enforced) and evidence (what’s logged).

Procurement note

If you want this to survive audits, insist on artifacts: requirements, evaluation gates, logs, incident procedures and reversibility clauses.