7 things worth seeing before your next round of research
Bite-size updates and learnings from the world of AI for customer insights work
AI Customer Research
How to use AI for Customer Research - and do it better than 99% - in <20 minutes monthly.
Subscribe:.rss.atom.json.md.m3u.pls
Live Last read · last published · next check
Latest episodes
How good is Fable really? And the intentional "poisoning" of your deep research
Benchmarking models and keeping track of the next threat to reliable insights
Your AI agent is wasting half its tokens — here's what to change
MCP tool descriptions have hidden instructions your agent follows literally. Five rules to override them.
The difficulty worth keeping when using AI
Desirable difficulties, cognitive offloading, and how to use AI without rotting your brain
How and why to create golden test sets
The core of a good AI insights experiment isn't your prompt. It's your test data.
The 5 levels of Claude Code for customer insights — and how to climb to the next
Where teams actually get stuck, and the smallest moves for you to level up quickly
Agents are now your primary readers
Your documents have two readers now - and AI will become the primary one. How to up your documentation game for your team's agentic POC.
Your Synthetic Users primer
Answering big questions for you before you're expected to at work.
Pressure-test your AI workflows across models, features in Lenny's and Aakash's newsletters, and a pile of AI news
Are you sure your AI workflows work as well as they could? A few ways to check.
Claude Code Skills, MCP apps, ChatGPT ads and more
Getting Claude to deliver 20x more—and improving your agents
Gemini 3 for video analysis
What works and doesn't, how to get the best out of it, and run your own rapid tests before end of year.
2025 Year In Review: The AI progress we've seen this year
AI Moderators, Synthetic Users, Video Analysis and more
4 hours of analysis + documentation in 14 minutes with zero uploads 👉 Claude Code for customer research.
The October 2025 Edition
Testing an AI-embedded browser for research uses, a 4-month roadmap to adopt AI meaningfully, and more...
The September 2025 Edition
The GPT-5 Issue | What’s better, what’s not, what to do with Agent Mode + a study on context windows.
The August 2025 Edition
Transcription evaluations head-to-head, and NotebookLM's "hidden" transcripts - are they any good? Plus a speedy test protocol...
The July 2025 Edition
Preventing AI's basic math errors, fast data cleaning, a viral study, a meta-prompting addition worth thinking about, and more...
The June 2025 Edition
A Claude 4 vs 3.7 reasoning test, new Claude system prompts, and "MCP" explained...
The May 2025 Edition
My new favorite LLM for customer research
Plus what we can use AI benchmark tests for, and scientists may have found the most efficient way to prompt...
AI designs: From insights to test-ready prototype views faster (a workflow)
Plus the tool options on the market, and data privacy issues -