AI tools for PMs, analysts, and knowledge workers — what actually saves hours, tested against the hype.
39 articles
We tested Perplexity and ChatGPT Search the way analysts actually use them: chasing every claim back to a source. Here's how their citation workflows differ and which one to trust.
Capture is solved; synthesis isn't. We compare the AI tools that actually turn scattered meeting notes, voice memos, and docs into a decision you can act on.
A step-by-step workflow for drafting product requirements documents with an LLM — what to feed it, what to keep human, and where AI-generated PRDs quietly drift off course.
We compared Granola, Fathom, and Otter on how they capture meetings, what they cost, and which workflow each one actually fits in 2026.
A practical comparison of NotebookLM and ChatGPT Projects for grounded research: source handling, citations, drift, and which one fits which job.
A new arXiv preprint proposes δ-mem, an online memory mechanism for LLM agents. What it claims, what remains unverified, and how to decide whether a persistent memory layer fits your agent, RAG pipeline, or chat app.
We tested Productboard's AI tools for surfacing themes and processing feedback. Here's where they save time on prioritization and where they quietly don't.
How to use an LLM to draft OKRs that survive scrutiny: forcing measurable key results, killing activity-disguised-as-outcome, and the prompts that catch vague goals.
A measured look at Maze's usability testing, surveys, and AI summarization features — what the AI layer actually does for product teams, and where you still need a human.
We tested Whimsical AI on the flowcharts and mind maps developers actually draw. What prompt-to-diagram does well, where it needs cleanup, and who it fits.
A measured look at Glean for product teams: how its permissions-aware enterprise search and grounded AI assistant work, who it fits, and why pricing is the catch.
AI fills the 'so that' clause of a user story with plausible reasons that were never in your research. Here is how to ground the prompt and keep an auditable why-trace.
A practical pipeline for clustering, tagging, and summarizing support tickets with LLMs so the patterns reach your product roadmap instead of dying in the queue.
A step-by-step workflow for running competitive teardowns with AI as a research assistant — without letting the model invent facts or flatten your judgment.
A practical workflow for turning raw git changelogs into release notes users finish reading — what AI does well, where it invents features, and the review gate you still need.
A hands-on look at Dovetail, the AI research repository for product teams: what its transcription, tagging, and insight features do well, and where a Notion-style stack still wins.
A practical look at where AI helps with product roadmapping in 2026 — drafting, summarizing, and prioritization — and where it still gets in the way.
A practical 2026 workflow for turning interview transcripts into themes with AI — where it saves hours, where it hallucinates, and how to keep your findings traceable.
A hands-on look at ChatPRD, the AI tool for drafting and critiquing product requirement docs. What it does well, where it falls short, and who should pay for it.
A hands-on look at Napkin AI, the text-to-diagram tool. What it generates, where it fits a product team's docs workflow, and the limits to know before you adopt it.
A hands-on look at Gamma's AI deck builder for product managers — what the generate-from-prompt flow does well, where it breaks down, and how its credit-based pricing works.
A 2026 reading list on data and metrics for product managers — picking the right metric, telling a story with it, setting goals that work, and running A/B tests you can trust.
A focused 2026 reading list on product discovery and strategy — the books that teach you how to find the right problem before you build, and how to choose what to build next.
A practical 2026 reading list for product managers — the foundational books on product thinking, discovery, strategy, and execution that working PMs keep recommending.
I lived in Superhuman, Shortwave, and Fyxer for several weeks each to see which AI actually clears an inbox. Here is how their triage, draft quality, and pricing compare for busy knowledge workers.
I spent two weeks moving real analyst work through Claude in Excel, Gemini in Google Sheets, and Rows AI. Here is how they compare on formulas, live data, trust, lock-in, and price.
A step-by-step, no-code guide for PMs and consultants to turn Claude into a personal knowledge assistant using Projects, custom instructions, and MCP connectors — with honest notes on limits and data hygiene.
I spent weeks running ChatGPT, Gemini, and Perplexity deep-research modes against real analyst tasks. Here is how depth, citations, speed, and cost actually stack up.
I spent weeks running both tools on the same research piles. NotebookLM wins on citation fidelity and ground-truth review; Claude Projects wins on synthesis and drafting. Here is how to choose.
We tested Perplexity Spaces, You.com, and Phind on real technical research workflows for two weeks. Here's which one wins for code, citations, and deep reports — and why most devs end up paying for two.
Testing how ChatGPT Projects, Claude Projects, and Gemini Gems retain instructions, files, and prior conversations across multiple sessions in 2026.
We ran the same five meetings through Granola, Fireflies, and Otter to compare transcription error rate, speaker attribution, and summary fidelity. Here is what each tool gets right and where it fails.
Two weeks running Reflect, Tana, and Capacities as a daily engineer's notebook. Capture latency, data portability, query depth, and where each one breaks down.
Three different bets on where company knowledge lives - Glean's connector-first enterprise search, Notion AI's workspace-native answers, and Mem's personal knowledge graph. Where each one actually fits.
I tested 6 'natural language to data' tools across the kind of queries PMs actually need to run — retention, funnel, cohort analysis. Some are genuinely usable now. Some produce confidently wrong answers.
Three meeting AI tools tested across 40+ product calls (discovery, internal sync, customer interviews). What each one is actually good at, where they all fail, and the per-seat math.
A product manager's honest review of Notion AI: where it replaces real PM work, where it produces convincing-but-useless output, and the workflow patterns that turned $10/month into hours saved per week.
How I use Perplexity Pro to do the competitive research a PM actually needs — pricing pulls, feature deltas, customer reviews — and the specific prompts that get past surface-level summaries.
The exact 4-prompt sequence I use to turn a 3-sentence feature brief into a reviewable PRD in 25 minutes — and the parts of the process Claude can't replace.
One email every Friday with the tools worth your time. No spam, unsubscribe anytime.