Library
13 published entries — scroll to load more.
Every diagnosis Agent Oracle has published, in one searchable place.
The library is the full archive of operating write-ups: pricing decisions, pipeline maths, hiring sequencing, churn post-mortems and the bottleneck patterns that show up again and again in small B2B companies. Each entry started as a real question an operator could not answer from a search engine, and each one ends with a claim specific enough to be wrong.
Use search when you already know the term you are chasing, and browse when you do not. Long-form pieces are marked with a reading time; shorter reference entries exist to define the vocabulary that the longer pieces lean on, so you can follow a chain of ideas instead of re-reading the same introduction.
Nothing here is sponsored and nothing is placed. Articles are drafted with AI research tooling against public sources, then edited, fact-checked and approved by the publisher before they appear in this archive. Where a claim depends on an assumption, the assumption is stated in the body rather than buried.
Geoffrey Moore is an American organizational theorist, management consultant, and author renowned for his work on technology adoption and market dynamics.
Michael Porter is an American academic known for his theories on economics, business strategy, and social problems. His work has significantly influenced both business management and public policy.
— end of library —
How to use the Agent Oracle archive
What is in here
The library is the complete published archive: pricing decisions, pipeline maths, hiring sequencing, churn post-mortems and the recurring bottleneck patterns of small B2B companies. Each piece began as a question an operator could not resolve from a search engine, and each one closes with a position specific enough that it could be shown to be wrong.
Reading times are shown so you can tell a reference definition from a long argument before you open it. Short entries exist to carry vocabulary that the long ones rely on, which is why chains of two or three pieces are common.
Search, browse, or follow a thread
Search when you already know the term you are chasing. Browse the categories when you know the symptom but not the name for it — revenue problems usually start in pricing and pipeline, effort-without-output problems usually start in process and hiring.
Every article links to the entries in the knowledge engine that define its terms, and back again, so you can move between argument and reference without losing your place.
Editorial standards
Nothing in the archive is sponsored, paid for or placed. Drafts are researched with AI tooling against public sources and then edited, fact-checked and approved by the publisher before they appear. Assumptions are stated in the body rather than buried, and corrections are published rather than silently patched.
How to get the most out of the archive
Three ways in
Search when you know the term. Browse the categories when you know the symptom but not its name — revenue problems usually begin in pricing or qualification, effort-without-output problems usually begin in process or sequencing.
Or read by date. The newest pieces respond to what operators are asking now, which makes the recent end of the archive a decent read on what is currently going wrong in small B2B companies.
Reading times mean something
A four-minute piece is a reference: a definition, a mechanism, a caveat. A twenty-minute piece is an argument with worked numbers, and skimming it will leave you with the conclusion and none of the conditions that make it valid.
Chains of two or three pieces are common because the long arguments deliberately lean on the short definitions rather than restating them badly.
Nothing here is a dead end
Every article links to the knowledge-engine entries that define its terms, and every entry links back to the arguments that use it. The full A–Z archive lists everything published, so no piece is reachable only through a recommendation algorithm.
Older pieces stay online. When one is superseded, it is updated or annotated rather than deleted, because a broken link is a worse outcome for a reader than a dated argument.