GeoIT: An Enterprise Architecture Framework in the Age of AI (AI-Native Enterprise Architecture Book 1) by Rogério Figurelli
English | December 9, 2025 | ISBN: N/A | ASIN: B0G5N5L4YC | 163 pages | EPUB | 1.66 Mb
In the age of large models and agentic AI, most enterprise architecture still thinks in straight lines: strategy on one end, implementation on the other.
GeoIT proposes a different picture. It treats enterprise IT as a set of loops that begin and end with stakeholders, and introduces Circles of Equivalence - explicit "rooms" where an initiative is allowed to change without losing its identity or its obligations. Instead of asking only "does this system work?", GeoIT asks "under what circle of promises is this system allowed to act, and when does a change mean we've launched something new?"
Building on the Geometry of Equivalence and the Circle of Equivalence theorem, the book translates clean mathematical ideas into practical tools for AI-intensive organizations. It introduces Fields (domains of meaning such as creditworthiness, consent, risk) and Overfields (cross-cutting obligations like fairness, privacy and safety) as the real vocabulary of architecture.
Around them, it defines four domain circles - Stakeholders, Business, Architecture, and Technology - that anchor a "Circle of Realization": a governed loop Stakeholders → Business → Architecture → Technology → Stakeholders. Stakeholders define the promises and obligations that must survive change; Business turns those promises into goals and operating constraints; Architecture binds them into structures that can be defended; Technology realizes them in running systems. Loops that genuinely close back to stakeholders are treated as "the same initiative"; loops that tear expose hidden holes in strategy, governance, or proof.
GeoIT then turns geometry into discipline through a small set of primitives: gates, thresholds, receipts and abstention. You'll see how to define problem, solution and implementation rooms for AI-intensive systems; how to decide which retrains, refactors and platform changes still count as equalities; how to encode fairness, safety and stakeholder rights as structural constraints, not just principles on a slide. A full part of the book is devoted to patterns and anti-patterns: how to start with a single honest circle, how to avoid "paper geometry," how to design safe agent corridors instead of free-roaming agents, and how to embed circles into CI/CD, MLOps and incident review.
Written for enterprise architects, AI and platform leaders, risk and governance teams, and regulators who need a sharper language for "the same system under change," GeoIT is both conceptual and practical. It won't ask you to throw away TOGAF or C4; it will sit above them as a geometric layer that makes your diagrams honest under AI-scale evolution.
If your organization is trying to scale AI without losing track of who it serves, what it promised, and where its systems are actually roaming, this book offers a way to draw - and keep - those boundaries on purpose.
Links are Interchangeable - Single Extraction
