



12 episodes
by LangChain · English
Welcome to Max Agency, a podcast about how the best AI agents are actually being built. Hosted by Harrison Chase, CEO of LangChain, each episode goes deep with the builders designing, deploying, and learning from real agent systems in the wild. From architecture decisions to evals, tooling, and failure modes, Max Agency is for people who want to understand what it really takes to build useful agents.









Clay’s new “Clay” agent rebuild for go-to-market, shifting from task-focused agents (e.g., Sculptor) to an end-to-end, outcome-oriented agent built on a coding-agent harness.
19 Nov 2026 · 1h 8m · 31 chapters
Unify’s AI agent platform for go-to-market (finding leads and drafting outreach) and how it cut AI agent costs 90–95% in two weeks by fixing prompt-caching and agent execution patterns.
13 Aug 2026 · 1h 8m · 29 chapters
Misaligned incentives in AI coding agents, shifting bottlenecks from training to evaluation, and how Cognition (Devon) scales proactive agents with better cost/quality via model routing and…
30 Jul 2026 · 50m · 19 chapters
Factory’s “software factory” for AI agents that turn enterprise “signals” (bugs, feedback, issues, telemetry) into validated deployed code, emphasizing model-independent harnesses that reduce cost…
16 Jul 2026 · 1h 18m · 36 chapters
LangChain product manager Ben Tannyhill explains LangSmith Engine, an “agent for agent engineers” that monitors production traces, clusters failures, proposes fixes (often via PRs), and generates…
2 Jul 2026 · 50m · 22 chapters
Sierra’s approach to building “agentic commerce” customer-experience agents that are simpler than typical agent harnesses: low-latency voice architecture, parallelized reasoning/listening/speaking,…
18 Jun 2026 · 1h 27m · 40 chapters
Benchling AI’s agent design for life-sciences R&D, emphasizing tool/context design, SQL-first data access, production-trace evals, and multi-model workflows to improve answer quality and speed.
4 Jun 2026 · 51m · 24 chapters
Cogent (Geng Sng) explains how Kojin/Cogent builds autonomous cyber-defense AI agents that must be correct every time—focusing on trust, context, safety, and fast remediation at “machine speed.”
22 May 2026 · 1h 15m · 33 chapters
Ramp’s “Ramp Sheets” agentic spreadsheet editor and the surrounding agent architecture, eval/testing, and self-improving monitoring loop; plus two R&D experiments (latent briefing with KV-cache…
7 May 2026 · 44m · 21 chapters
How Listen’s CTO describes building an “agent-first” system for analyzing hundreds of multimodal user interviews/surveys/focus groups using a table-like data model, subagents, and quality/evaluation…
23 Apr 2026 · 48m · 26 chapters
How Hex (podcast guest Izzy Miller, AI engineer) builds “data agents” that reason like human data analysts, emphasizing context harvesting, tool orchestration, and evaluation/verification.
9 Apr 2026 · 1h 8m · 25 chapters
Intro episode for “Max Agency,” a podcast focused on how top agents are built by “builders like you.” Guest(s): No guests are mentioned in the provided transcript.
8 Apr 2026 · 1m