01 · DOMAIN
A bounded business area
Onboarding, lending, claims, risk — anything with its own vocabulary, stakeholders and stories. Each domain is a tenant in the graph.
Decision Atlas maps every domain into use cases, every use case into artifacts — BRDs, stories, code, transactional data, incidents — each with its own graph. Query Assist runs across all of them and returns one answer.
Decision Atlas organizes knowledge in three nested rings — each level enriches the next, and every link survives as a queryable edge in the final graph.
Onboarding, lending, claims, risk — anything with its own vocabulary, stakeholders and stories. Each domain is a tenant in the graph.
Each domain holds many use cases — onboard SME, approve loan, detect anomaly. They define the artifacts you'll need to capture.
Each artifact — BRD, story, code, transactional data, incident — builds its own knowledge graph. Entities, relationships and provenance, all queryable.
Knowledge Assist routes a natural-language question across every artifact graph in the domain — incidents, code, stories, BRDs — and returns one unified answer.
Decision Atlas builds a dedicated knowledge graph for every artifact a use case produces. Query Assist runs across all of them — every artifact, every use case, the whole domain — and returns one answer.
Business requirements parsed into intent, scope, actors and acceptance criteria.
DocumentUser stories generated from BRDs — split, linked to acceptance criteria and tickets.
GeneratedRepos ingested via ZIP or GitHub. Modules, functions and dependencies become nodes.
ZIP · GitHubDatabase schemas and live transactions feed the graph with real-world cardinality.
DB · StreamPost-mortems and on-call logs are linked back to the code and stories they touched.
Log · PIREvery artifact contributes nodes and edges. Hover, query, drill — the graph is what your domain actually looks like under the hood.
From an empty domain to a queryable knowledge graph in minutes.
Spin up a tenant for lending, claims, onboarding — whatever the bounded context is.
Approve loan, onboard SME, detect anomaly — each becomes a leaf in the graph.
Drop BRDs, push code, point at a DB. Decision Atlas parses and links them automatically.
Ask in natural language. Trace impact, surface gaps, generate new artifacts on demand.
Start with a use case, ingest one artifact, and watch the graph form in real time.