EPISODE · May 5, 2026 · 36 MIN
MCP Isn’t Always the Answer: 7 Experiments Building AI Agent Interfaces at PostHog
from Podcast de Itnig: Historias de startups · host itnig
Georgiy Tarasov, AI Product Engineer at PostHog, shares what PostHog learned from seven experiments building interfaces for AI agents — and why MCP isn’t always the right answer. As customers started asking for “all PostHog AI features in my agent,” PostHog had to rethink how its product should work outside the browser: inside Claude Code, Cursor, Codex, CLIs, MCP clients, and local agent workflows. In this talk, Georgiy walks through PostHog’s experiments with handwritten agent tools, traditional MCP, CLI-first interfaces, MCP with a CLI-like shape, code execution, dynamic toolsets, and SQL-based retrieval. He compares the trade-offs across developer experience, context bloat, tool discovery, token efficiency, latency, eval results, and real customer usage. The core lesson: your new user might not be a human in a browser tab — it might be your customer’s AI agent.👉 Subscribe to Itnig for more conversations about real business, startups and brands.🎙️ Want to join the Itnig podcast or sponsor one of our episodes? Appear on the podcast: https://tally.so/r/wo1Poe Sponsor the podcast: https://tally.so/r/3EERLNABOUT ITNIG: 🐦 X - https://x.com/itnig 💡 LinkedIn - https://linkedin.com/company/itnig 📸 Instagram - https://instagram.com/itnig 💌 Newsletter - https://itnig.net/newsletter/ 🌐 Web - https://itnig.net/ LISTEN TO OUR PODCAST ON: 🔊 Spotify: http://bit.ly/itnigspotify 🎙️ Apple Podcast: http://bit.ly/itnigapple00:00:00 Intro & welcome — AI Builders BCN first edition00:01:04 Georgiy introduces himself & PostHog00:02:14 The challenge: shipping AI agents on a complex product00:04:37 First MCP experiments & why they didn't scale00:05:26 Token optimization problems & early lessons00:07:08 Rethinking the approach: sandboxing & unified interfaces00:08:13 Deep dive: how MCP prompts & instructions really work00:10:26 Optimizing MCP tools for different model providers00:12:55 Experiment #1 — grouping tools by intent00:14:15 Switching to a resource-based approach00:18:24 Advantages & limitations of the code execution approach00:22:30 MCP with native code execution — Georgiy's favourite00:25:15 How GitHub optimizes their MCP (grouping by intent)00:25:43 Agent skills: what they are and why they matter00:26:43 Building skills with regex, access controls & data00:27:52 Benchmarking: testing with real vs synthetic data00:29:38 Results & what actually worked in production00:31:35 Key takeaways & closing thoughts00:31:43 Q&A 00:36:29 End Recorded at AI Builders Barcelona. Speaker: Georgiy Tarasov, AI Product Engineer at PostHog Topics: MCP, AI agents, Claude Code, Cursor, Codex, CLI, codegen, agent interfaces, developer tools, PostHog, AI engineering
What this episode covers
Georgiy Tarasov, AI Product Engineer at PostHog, shares what PostHog learned from seven experiments building interfaces for AI agents — and why MCP isn’t always the right answer. As customers started asking for “all PostHog AI features in my agent,” PostHog had to rethink how its product should work outside the browser: inside Claude Code, Cursor, Codex, CLIs, MCP clients, and local agent workflows. In this talk, Georgiy walks through PostHog’s experiments with handwritten agent tools, traditional MCP, CLI-first interfaces, MCP with a CLI-like shape, code execution, dynamic toolsets, and SQL-based retrieval. He compares the trade-offs across developer experience, context bloat, tool discovery, token efficiency, latency, eval results, and real customer usage. The core lesson: your new user might not be a human in a browser tab — it might be your customer’s AI agent.👉 Subscribe to Itnig for more conversations about real business, startups and brands.🎙️ Want to join the Itnig podcast or sponsor one of our episodes? Appear on the podcast: https://tally.so/r/wo1Poe Sponsor the podcast: https://tally.so/r/3EERLNABOUT ITNIG: 🐦 X - https://x.com/itnig 💡 LinkedIn - https://linkedin.com/company/itnig 📸 Instagram - https://instagram.com/itnig 💌 Newsletter - https://itnig.net/newsletter/ 🌐 Web - https://itnig.net/ LISTEN TO OUR PODCAST ON: 🔊 Spotify: http://bit.ly/itnigspotify 🎙️ Apple Podcast: http://bit.ly/itnigapple00:00:00 Intro & welcome — AI Builders BCN first edition00:01:04 Georgiy introduces himself & PostHog00:02:14 The challenge: shipping AI agents on a complex product00:04:37 First MCP experiments & why they didn't scale00:05:26 Token optimization problems & early lessons00:07:08 Rethinking the approach: sandboxing & unified interfaces00:08:13 Deep dive: how MCP prompts & instructions really work00:10:26 Optimizing MCP tools for different model providers00:12:55 Experiment #1 — grouping tools by intent00:14:15 Switching to a resource-based approach00:18:24 Advantages & limitations of the code execution approach00:22:30 MCP with native code execution — Georgiy's favourite00:25:15 How GitHub optimizes their MCP (grouping by intent)00:25:43 Agent skills: what they are and why they matter00:26:43 Building skills with regex, access controls & data00:27:52 Benchmarking: testing with real vs synthetic data00:29:38 Results & what actually worked in production00:31:35 Key takeaways & closing thoughts00:31:43 Q&A 00:36:29 End Recorded at AI Builders Barcelona. Speaker: Georgiy Tarasov, AI Product Engineer at PostHog Topics: MCP, AI agents, Claude Code, Cursor, Codex, CLI, codegen, agent interfaces, developer tools, PostHog, AI engineering
NOW PLAYING
MCP Isn’t Always the Answer: 7 Experiments Building AI Agent Interfaces at PostHog
No transcript for this episode yet
Similar Episodes
Mar 26, 2026 ·1m
Jan 2, 2026 ·47m
Dec 21, 2025 ·46m