Aggregation
Compacts raw rows into gold, analysis-ready records before anything else runs.
Capabilities / selected per problem
Each engine has one clear job.
ADE composes typed, deterministic capabilities instead of forcing one model onto every dataset. Every output keeps its pattern, reasoning, and confidence attached.
Compression
Reduces noisy or repetitive signals while keeping the patterns that matter.
Clustering
Deterministic K-Means groups similar behaviour so outliers stand out.
Regression & forecasting
Projects demand, revenue, and risk reproducibly on every run.
Fuzzy scoring
Produces readable low, medium, and high levels with fired-rule traceability.
Scenario engine
Compares alternative decisions side by side before your team commits.
Confidence layer
Shows how much to trust a result and what still needs human review.
Recommendation engine
Generates structured actions with priority, reason, and confidence.
Decision lifecycle
Tracks recommendations from suggested to approved, executed, and measured.
Advanced and optional
Opt into depth. Never pay for theatre.
Advanced capabilities remain explicit and contract-driven. They do not silently enter the default decision path.
Genetic optimisation
Searches for stronger parameter configurations with deterministic seeds.
Plugin SDK
Lets domain verticals plug into the same typed runtime.
Tenant isolation & roles
Protects every project with row-level security and scoped roles.