The news: The second annual AI Agent Conference in New York City this week revealed business focus is quickly shifting from agentic adoption to how to govern, measure, and monetize agents at scale.

Three interconnected themes emerged: the collapse of agent trust infrastructure as a standalone crisis, agentic commerce rising as a customer-facing layer, and the need for evaluation as a prerequisite to enterprise adoption.

Zooming in: Representatives from OpenAI, Microsoft, Amazon Web Services (AWS), Meta, Bright Data, and more focused on three top priorities.

1. Trust and identity are the new technical bottleneck.

• As multi-agent systems grow inside enterprises, data security and AI agent autonomy become difficult to reliably track. 
• Businesses are finding a “trust paradox”—agents fast enough to create value are also fast enough to create liability. 
• Security providers at the conference coalesced around agent identity as a standalone market category, distinct from both traditional Identity and Access Management (IAM) and large language model (LLM) safety.

2. Agentic commerce is rewriting the customer journey.

• Sessions from PayPal, Lightspeed Commerce, Groupon, and Chief Digital Agency treated agentic commerce as a first-class vertical, not an experiment.
• As AI agents help drive discovery, complete purchases, and handle service resolution for consumers, the traditional impression-to-conversion funnel breaks down entirely.
• Brands are angling for agent preference and positioning, not clicks. This represents an emerging form of search and discovery optimization, one that martech and ad tech stacks, designed for human attention, aren’t built to address.

3. Evaluation and observability are the path to enterprise adoption.

• Galileo, Microsoft, ADP, Meta, and DataRobot all confirmed a universal AI adoption speed bump—enterprises won’t scale up their AI agent deployments unless they have repeatable, reliable ways to evaluate them.
• Nearly three-quarters of companies have yet to generate meaningful value from AI, with many still stuck in pilot phases despite growing investment, per The World Economic Forum.
• Most organizations don’t have the measurement infrastructure in place to confidently move from pilot to production.

Recommendations for brands and agencies: Many industries are facing the same challenges in adopting agentic solutions, and all need to start somewhere. 

• Watch the agent identity space closely. Agentic identity management will determine which enterprise AI deployments scale and which stall. Governance infrastructure, like vendor due‑diligence requirements, standardized approval workflows, clear data‑use and security policies, and tools to audit models, should be procurement prerequisites.
• Audit your discoverability for AI agents. As autonomous agents handle more product discovery and purchase decisions on behalf of consumers, brands that aren't optimized for agent-readable content risk being invisible.
• Reframe creative production ROI. Half of worldwide marketers are using AI for text, images, and videos, per Canva, yet most brands still lack a reliable way to measure what that investment is actually returning. The tools have become mainstream, but the evaluation frameworks still need to catch up.

This content is part of EMARKETER’s subscription Briefings, where we pair daily updates with data and analysis from forecasts and research reports. Our Briefings prepare you to start your day informed, to provide critical insights in an important meeting, and to understand the context of what’s happening in your industry. Non-clients can click here to get a demo of our full platform and coverage.