If you block an AI agent, you are teaching it to recommend someone else

Encore’s scan of more than 1,500 organizations across 25 industries found that, as of Oct. 23, 94% of sites could not accept an AI-agent connection, 26% blocked agents that identified themselves, and only 1% offered customers a way to authorize a limited, revocable agent task. The commentary argues that indiscriminate blocking can make businesses invisible to customers’ AI assistants, while only 6% of scanned organizations had published a position on AI agents. It recommends machine-readable access rules and limited, direct agent connections rather than simply allowing all automated traffic.

Analysis

The investable signal is not the article’s 94% readiness statistic; it is a possible shift in who controls digital customer discovery. If assistants become a meaningful transaction channel, businesses that expose reliable, permissioned APIs could gain completed transactions, while friction-heavy sites risk losing consideration without seeing conventional funnel abandonment. Over 6–18 months, that could weaken some travel, insurance and financial-service firms’ control over distribution and increase the bargaining power of agent platforms. It may also pressure paid-search economics if agents substitute direct task completion for browsing and comparison.

The near-term counterweight is fraud, privacy and liability: authenticated agents acting inside accounts could raise losses and force banks and merchants to invest in identity, delegated authorization and monitoring. That creates demand for cybersecurity and identity infrastructure, but may also benefit existing bot defenses if they evolve rather than disappear. The article is an Encore-sponsored argument, and its public-page scan does not establish agent-driven transaction volumes, lost revenue, or that the sample represents commercial activity. “Readiness” is not evidence of earnings impact.

Days: limited basis for a directional trade. Over 1–3 months, watch for major platforms or merchants adopting revocable, scoped agent authorization and for disclosure of agent-originated transactions. Over 6–18 months, the key test is whether agents drive measurable conversion and distribution-cost changes. The contrarian point: permissive agent access is not automatically a competitive advantage; trusted authorization and fraud controls may matter more than access itself.

Key Decisions for Investors

• No immediate thematic position on the readiness statistic alone; treat it as a vendor-sponsored adoption signal, not proof of demand or lost sales.
• Add cybersecurity and identity providers with delegated-authorization, bot-risk and account-protection capabilities to the watchlist; do not assume legacy bot-management vendors lose if controls adapt.
• Track travel, insurance and financial-service firms for agent-related conversion, referral mix, customer-acquisition cost and fraud-loss disclosures. Consider relative shorts only if distribution economics deteriorate without offsetting lower acquisition costs.
• Reassess after evidence of production deployments, scoped/revocable permissions and measurable agent-sourced transactions. Falsifiers include persistently low agent usage, rising fraud losses that prompt broad blocking, or no reported change in conversion and acquisition costs over coming quarters.