Links
Home
Info
I help organizations deliver production AI chatbots, AI-ready and localization-ready documentation, and Gartner-grade submissions across global product lines.
Most people outside enterprise software have never heard of Gartner or the Magic Quadrant, and most people inside it underestimate how much a submission depends on documentation discipline rather than marketing polish. Here is what it actually is, why the stakes are as high as they are, and why I treat it as an AI readiness problem.
Gartner is a research and advisory firm that large enterprises pay for guidance on which vendors and products to trust. The Magic Quadrant is their signature report format. For a given category, such as backup software, cloud platforms, or CRM, Gartner evaluates every major vendor on two axes: completeness of vision and ability to execute, and places each one into a quadrant: Leaders, Challengers, Visionaries, or Niche Players.
That chart carries an enormous amount of weight. Enterprise procurement teams use it directly to build vendor shortlists, and being named a Leader functions as a trust signal that a lot of buyers treat as close to decisive. A strong placement gets a company into deals it would never have been considered for otherwise. Companies that land well in the Magic Quadrant can genuinely print money off that placement alone. Companies that get placed poorly, or dropped from the report entirely, send the opposite signal to the exact buyers who were about to write a check.
What actually gets submitted to Gartner is an evidence package: detailed answers to capability questionnaires, product documentation, customer references, and competitive positioning narratives, all of it expected to hold up under an analyst's follow up questions. The quality bar here is not the quality bar for marketing copy. Every claim needs to be accurate, current, and defensible, because analysts do follow up, and because the customer references listed in the submission get called and asked to confirm what was written about them.
This is the same principle behind everything else I do with documentation and AI: the documentation is the fuel, and the system around it determines whether that fuel is trustworthy. For a Gartner submission, that means building the narrative only from approved product documentation and vetted knowledge content, never from marketing assumptions or content that has gone stale. Every claim goes through governance and provenance checks before it goes into the submission, so there is a real answer to where it came from, not just confidence that it sounds right. If AI helped draft any part of the narrative, that draft gets checked against the same provenance chain as everything else. A submission is exactly the kind of document a professional analyst is going to fact check, so nothing unverifiable can be allowed to slip into it.
When an analyst asks how you know that, you need a real answer, not a good sentence.
Here is a nuance most people get wrong: a large language model does not need a rigid taxonomy nearly as much as people assume. Technically, that's because embeddings capture semantic relationships directly from the content and its context, so a model can relate two topics to each other without an explicit category tree telling it they're related. It generalizes from the text itself.
Taxonomy still matters a great deal in this work, just for two different reasons. First, RAG needs deterministic filtering, not just semantic similarity. Without taxonomy tags for product and version, a retrieval system can quietly pull a capability description from the wrong release or the wrong product line into a submission answer, and a wrong product version is exactly the kind of error that gets a claim challenged. Second, and just as important, is the human side. The analyst relations team, the documentation team, and the reviewers checking a submission before it goes out are not language models. They need a taxonomy so they know where new evidence gets filed, where to look when a question comes in, and how to audit what has already been submitted. Taxonomy keeps the human organization around the evidence library coherent, even in the places where the model itself could technically get by without it.
The teams that do well in the Magic Quadrant are not the ones with the best marketing language. They are the ones whose documentation was already governed, sourced, and organized before the submission clock started, so building the evidence package was a matter of assembling verified content instead of scrambling to justify claims nobody could trace back to anything real.
I help teams build the governed evidence library and provenance chain that makes a submission defensible.
Get in touchI help organizations deliver production AI chatbots, AI-ready and localization-ready documentation, and Gartner-grade submissions across global product lines.