This insight summarizes the August 2026 MarTech piece "What it takes for AI to find, trust, and act on your brand" by Benu Aggarwal, founder and president of Milestone Inc.
Traditional search rankings no longer measure what matters. Whether AI systems discover, understand, recommend, or transact on your brand is a different question — and most brand analytics don't answer it. Aggarwal calls this the "measurement gap" and argues brands now need to optimize for machine comprehension alongside human navigation.
The journey itself has changed. Traditional search: query → ranking → click → website → decision. AI search: intent → research → retrieval → synthesis → recommendation → action. With bots now representing 57.5% of all web traffic, your site is serving two audiences at once, and AI engines break single queries into many sub-searches — comparisons, reviews, alternatives — before synthesizing an answer. They favor information density over content length.
A three-layer optimization framework:
- Eligibility layer. Crawlability, structured data, and machine-readable content. If AI can't access or extract your information cleanly, nothing downstream matters.
- Recommendation layer. Trust built through consistent data across platforms, corroboration from trusted third-party sources, freshness, and complete answers. AI won't recommend what it can't verify.
- Transaction layer. Live APIs, authentication protocols, and machine-executable transaction capabilities so an AI agent can actually act — book a room, buy a product — on the brand's behalf.
What to do now. (1) Treat schema and entity architecture as foundational infrastructure, not a checkbox. (2) Enforce consistent information across website, structured data, third-party listings, and directories — divergence kills trust. (3) Monitor the mention-to-citation gap: frequent brand mentions without citations point to an accessibility or trustworthiness problem. (4) Build measurement systems tracking presence (citations, recommendations), readiness (why you're winning or losing), and business impact. (5) Move toward agent-ready architecture that supports emerging protocols rather than optimizing for any single standard. (6) Accept that recommendations can happen without a website visit — success metrics must expand beyond click-through.