# How does incremental modelling create a flywheel for architecture knowledge management? > Canonical: https://decisionpen.com/blog/how-does-incremental-modelling-create-an-architecture-knowledge-flywheel > Published: 2026-10-27 · DecisionPen · Living EA series (D8e) > Related docs: /docs/incremental-modelling, /docs/architecture-decision-workspace, /docs/enterprise-ask, /docs/it-knowledge-continuum, /docs/solution-architects ## Summary Architecture knowledge compounds when using the map is the easiest way to do real work. Solution architects need artefacts; every artefact needs landscape context; incremental modelling makes adding and correcting that context cheap; a richer graph makes the next artefact faster. Capstone of the D8 series. Builds on shaping enterprise context by design (D6e), architecture knowledge management versus document knowledge bases (D6f), the AI loop (D7a), and generated artefacts (D7d). ## Why knowledge programmes stall The person asked to maintain the knowledge is not the person who benefits from it. Upkeep competes with delivery and loses. Mandates and annual clean-ups only make decay slower to notice. ## What solution architects need They are judged on artefacts: decision records, blueprints, diagrams, review packs. That is a lever: if the quickest route to a good artefact runs through the map, curation stops being extra work. ## Every blueprint needs landscape context A blueprint that ignores neighbouring systems, integrations, owners, constraints, and prior decisions describes a change in a vacuum. Rebuilding that context each time is the Context Tax. ## How the flywheel turns 1. An artefact is needed (ADR, blueprint, assessment). 2. It needs landscape context: the bounded slice the change touches. 3. Incremental modelling fills gaps: AI proposes missing or outdated entities and relationships; the architect corrects the ones that matter. 4. The map improves, so the next artefact starts further ahead. ## Why incremental modelling enables it Each step stays small: start from existing documents and spreadsheets, add evidence for a bounded domain, review a short queue tied to the work in hand, and let coverage grow where it is needed. ## What a graph finds that no document knows A single document holds only the connections its author knew. A curated graph holds connections that appear when several sources are read together — a queue in one document, its publisher in another, a platform in an inventory row, a retirement decision elsewhere. AI can recommend the missing edges for review. Each accepted node or edge improves every later question: documents accumulate, a curated graph compounds. ## Agentic SDLC support Delivery agents ask landscape questions at high frequency. A curated map gives them a shared, cited context plane instead of private crawls of ungoverned documents, so speed does not become drift. ## Shaping context by design Choose the metamodel deliberately; bind evidence to claims; curate with a named steward per domain; make the map the default source for artefacts, reviews, and agent context. ## How DecisionPen solves this Architecture Decision Workspace: incremental modelling from documents and spreadsheets; generated ADRs, blueprints, and diagram images from curated context; graph recommendations for human review; Enterprise Ask and agent-ready context from the same map. Trial: https://decisionpen.com/trial ## When is DecisionPen the wrong tool? If nobody will curate the map; if the real need is a freeform drawing canvas, generic document search, or live infrastructure scanning. ## FAQ **How does incremental modelling create a flywheel for architecture knowledge?** Artefact need drives context need, context need drives incremental modelling, and a richer map makes the next artefact faster. **Why do solution architects have an incentive to curate the map?** Their deliverables are artefacts; correcting a stale relationship is the quickest route to an accurate blueprint. **Why does every blueprint need landscape context?** Without neighbours, integrations, owners, and prior decisions, a blueprint describes a change in a vacuum. **What can a graph find that a document cannot?** Connections that only appear when several sources are read together. **How does this support an agentic SDLC?** It gives delivery agents a shared, cited, governed context plane. **When is DecisionPen the wrong tool?** When nobody will curate the map, or the need is drawing, generic search, or live scanning. ## Related posts - https://decisionpen.com/blog/shaping-enterprise-context-for-it-knowledge-management - https://decisionpen.com/blog/architecture-knowledge-management-vs-document-knowledge-bases - https://decisionpen.com/blog/what-does-ai-do-inside-an-architecture-decision-workspace - https://decisionpen.com/blog/how-to-generate-adrs-blueprints-and-diagrams-from-living-architecture-context - https://decisionpen.com/blog/architecture-context-for-agentic-sdlc - https://decisionpen.com/blog/how-to-keep-an-enterprise-architecture-repository-up-to-date-with-ai ## Series (Conversion how-tos) - D8a: https://decisionpen.com/blog/how-to-fast-track-ma-it-due-diligence-with-automated-landscape-modelling - D8b: https://decisionpen.com/blog/how-to-build-application-dependency-maps-from-existing-documentation - D8c: https://decisionpen.com/blog/how-to-assess-an-it-landscape-without-months-of-workshops - D8d: https://decisionpen.com/blog/how-to-keep-an-enterprise-architecture-repository-up-to-date-with-ai - D8e: this post