Decide on evidence. Read like it, too.
Analysis of the competitive-intelligence trade by Glenn Hurn: the legacy analyst model, generative-AI research shortcuts, and what independent evidence can actually verify. Every claim carries its source.
A Chat Window Is Not a CI Program: The Evidence-Hierarchy Gap
Why prompting a copilot is not competitive intelligence: five failure axes, the published evidence on chatbot reliability, and what a real CI program requires.
Hallucination Is Not a Bug You Can Prompt Away
OpenAI's own research says hallucination is structural, not a prompt problem. What that means for procurement decisions built on generative output.
Confidently Wrong: What the AI-Search Citation Studies Actually Found
Independent studies keep measuring the same thing: AI search answers with confidence and cites with error. What that means for competitive research.
80% of Features Are Rarely or Never Used. Opinion-Driven Roadmaps Are Why.
Pendo found 80% of software features are rarely or never used. That is not an engineering failure. It is what opinion-driven roadmaps produce.
Six to Ten People Buy Your Product. You're Only Arguing With One of Them.
Gartner counts six to 10 stakeholders in a typical B2B deal, each with independent sources. Why provable claims beat persuasive ones.
65% of Your Deals Are Competitive. Your CI Program Is One Person and a Wiki.
Crayon's data says 65% of software deals are contested, yet most companies run CI with two or fewer people. The math of that mismatch, and how to close it.
Who Does the Analyst Work For? The Dual-Customer Problem in Research
Research firms sell to the vendors they rank and to the buyers who trust the rankings. Courts cleared Gartner. The structure still deserves scrutiny.
The Quadrant Lag: Annual Snapshots in a Weekly Market
Analyst quadrant data is collected months before publication and read for a year after. The staleness arithmetic, and what it costs vendor shortlists.
The Courts Called It Opinion. Your Procurement Process Treats It as Proof.
Two vendors sued Gartner over Magic Quadrant placements. Both courts called the quadrant protected opinion. Your decision record should weight it that way.
The $6.5 Billion Opinion: What a Legacy Analyst Subscription Actually Buys
Gartner booked $6.5 billion in 2025 selling research that courts classify as opinion. What a subscription buys, and what evidence-class intelligence looks like.