The open questions are legal, not technical. Mack asks whether an agent can accept a liability cap tied to twelve months of fees, agree to automatic renewal, or allow a supplier to use company data. She argues that interoperability cannot answer these questions, because an agent may be able to complete a transaction while lacking authority to make the commitments inside it. Obligations also continue after signing, as agents may need to approve invoices, enforce service credits, or decide on renewals. Her own commentary is the only source here, and it presents her view, not established findings. She writes from the perspective of someone who runs a legal technology company.
For counsel, the practical takeaway is to engage before these systems are designed, not after deployment. Mack recommends replacing generic playbook language such as "commercially reasonable" with structured positions, thresholds, and escalation rules that an agent can actually apply. The column sits within a wider debate. Deloitte notes that U.S. e-commerce law assumes human assent, while UETA Section 14 allows contracts to form through electronic agents without human review. A Lexology analysis of merchant platform agreements found that risk falls mostly on merchants, with platform liability capped and merchant indemnities broad. The UK Information Commissioner's Office warned in January 2026 that explicit consent may be hard to obtain in agent deployments. No comprehensive U.S. statute on platform contracting is expected soon, so the contract terms companies negotiate now will set the rules.