As AI agents become autonomous transactors (capable of browsing, purchasing, and entering into commitments on behalf of their users), the trust infrastructure meant to govern their transactions is being built around a blind spot. Interoperability protocols, industry initiatives and academic work focus almost exclusively on agent verification (Know Your Agent: its identity, its authorisation, its attachment to a principal), without ever examining the destination towards which the agent is heading. This blind spot proves all the more critical given that AI-driven fraud is growing rapidly and exploits precisely what the agent lacks: the contextual judgement that constitutes the human buyer's first line of defence. This paper sets out to fill that gap by defining a distinct conceptual category, Know Your Destination (KYD): a pre-action verification framework that assesses the security, legal compliance, behavioural integrity and public trust of a destination before any commitment by the agent. After mapping the agentic commerce ecosystem and the legal framework applicable to it, the paper sets out the architecture of KYD and argues that it constitutes a requirement whose absence, as agentic transactions reach scale, becomes a systemic risk.
Bedar et al. (Mon,) studied this question.