Your LLM's Web Search Is Expensive and Slow. Here's How to Test It.
​📅 July 28, 2026 · 10a PT · Live on Zoom
​The Problem
​Does the agent you built for your customers have access to the web? Every team building AI agents has bolted an LLM onto a search API and called it done. Nobody's checked whether that setup is actually working.
​Here's the math: an org running 1,500 people, half of them hitting web search 10x a day, burns through 5.5 million queries a year — most of it running on infrastructure that was never built for how agents actually search.
​To put it in layman's terms: your built-in web search is Walmart. Indexing is Chanel.
​This session shows you the difference, live.
​What You'll Walk Away With
​What's actually happening under the hood of your LLM's web search call - SERP wrapper, scraper, or real index, and why that distinction changes everything downstream
​Why it's costing you tokens and accuracy - the API vs. Index breakdown, with real benchmark data instead of vendor claims
​A live side-by-side test - same query set, naive setup vs. a real index, so you can see the gap instead of taking it on faith
​A self-test framework you can run on your own stack the same day
​Who This Is For
​AI engineers, ML engineers, and technical PMs building or evaluating the retrieval layer of an AI product - especially if your team is running agentic workflows at any real volume.
​Your Host
​Antonio Mallia is the Founder and CEO of Seltz, building web retrieval infrastructure for AI agents. He built web search for Alexa, previously worked on Amazon's AGI team, Pinecone, and Bloomberg, and holds published research in information retrieval at scale.
​About Seltz
​Seltz builds web retrieval infrastructure for AI agents, a fast, real-time index designed for machine consumption. Learn more about Seltz at seltz.ai
​About AlphaSignal
​AlphaSignal is one of the most-read AI newsletters in the world, trusted by 300,000+ subscribers - developers, researchers, engineers, and AI practitioners who want to stay ahead of what's happening in artificial intelligence.
​We cover model releases, research papers, new tools, and practical tutorials written for people who actually build with AI, not just read about it.
​Subscribe today at AlphaSignal.ai