We break down what changes when AI agents become a bigger source of web visits than humans, and why measurement is the only sane response to generative search volatility. We also map the new playbook for go-to-market leaders who want durable advantage by building context, instrumentation, and systems that actually learn.
•Agent analytics and deeper web measurement through server logs and segmentation
•Turning GEO and AEO into a trackable marketing channel with baselines and experiments
•Why the website shifts from destination to structured knowledge layer for AI agents
•Context engineering through schema, FAQs, and connected site structure
•Why “renting cognitive logic” from LLMs fails to create a moat
•Decision tracing as a system for learning from GTM experiments over time
•AI integration tax in fragmented stacks and how dynamic blindness breaks workflows
•What a harness is and how constraints, monitoring, and stopping rules reduce risk
•Leading the shift from point-solution operator to multi-agent orchestrator
•Moving boards from token costs to value per token and closed-loop outcomes
AI agents are quietly rewriting your analytics dashboard, and most teams are still looking at the old numbers. When bots and crawlers can drive more web visits than humans, “traffic” stops being a simple KPI and becomes a strategy problem: Who is visiting, what model sent them, where do they land, and do they convert? We sit down with Alexander Liss, Executive Advisor of AI Transformation at Brainworks and former VP of Data Science and AI at Huge, to talk about agent analytics, server logs, segmentation, and how to treat generative search optimization as a real, measurable marketing channel.
We also go into the "website’s" near-death and rebirth. The site is not disappearing, but early discovery is moving into AI Overviews and assistants, which means your content has to work as structured knowledge for machines and as high-trust depth for humans who arrive later. We cover context engineering, schema, FAQ patterns, and why social listening is back as models pull signals from places like YouTube, Wikipedia, and community forums, then change their minds a week later.
From there, we get blunt about competitive advantage. If you are only renting cognitive logic from base models from Anthropic, OpenAI and Google, you are not building a moat. We unpack decision tracing, the AI integration tax, dynamic blindness in multi-agent workflows, and what it means to build a harness with constraints, monitoring, and stopping conditions so AI systems stay useful and cost-effective. We close with a fast, fun Spark Tank segment on Japanese business systems plus the “three feet from gold” resilience story. Subscribe, share with a growth leader, and leave a review. What part of your go-to-market stack needs better AI measurement first?
Alexander Liss: https://www.linkedin.com/in/aliss77777
Alexander Liss is the Executive Advisor of AI Transformation at BrainWorks and former VP of Data Science & AI at Huge. A data leader and systems-builder based in Denver, Colorado, his recent work includes pioneering Huge’s emerging GEO practice and deploying NBCUniversal’s award-winning conversational assistant, Oli, for the 2026 Winter Olympics. His prior leadership spans scaling enterprise analytics and machine learning strategies at global digital powerhouses like Accenture Song, VML and 22squared.
He is currently pursuing his Master of Science in Artificial Intelligence from the Georgia Institute of Technology, holds an MBA from NYU Stern School of Business, and earned his Bachelor's in Japanese Language and Literature from The George Washington University.
Website: https://www.position2.com/podcast/
Rajiv Parikh: https://www.linkedin.com/in/rajivparikh/
Email us with any feedback for the show: sparkofages.podcast@position2.com
