Kindynos EVA provides event-driven model orchestration across existing analytical models — spreadsheets, internal libraries, vendor engines, APIs and services. It connects real-world Events to the appropriate models and carries their results into a governed consequence chain with identity, provenance and reproducibility preserved.
The model remains responsible for the calculation it performs. EVA determines when that capability is relevant, what governed state should be supplied, how results connect to other consequences, and how the complete analytical path is recorded.
Identify the occurrence, affected system and evidence that matters.
Map the Event into approved parameter or state changes.
Choose one or several governed analytical providers.
The spreadsheet, library, service or vendor engine performs its specialist calculation.
Carry outputs into assets, stakeholders, further models, provenance and decisions.
A model does not need to be rewritten into a Kindynos format to participate. The integration can match the way the model already operates.
Existing Excel and analyst spreadsheets can remain the analytical asset. EVA can connect spreadsheet models to an Event-driven workflow, govern controllable inputs, recalculate a working copy, read defined outputs and preserve the resulting state and provenance.
Python, R, C++, Java and proprietary quantitative libraries can be reached through typed adapters while remaining under the institution's own change-management process.
Specialist commercial technology can be invoked through an API, SDK, command line, local service or controlled wrapper without exposing proprietary model internals.
Models can run as private APIs, local executables, containers, on-premise calculation services or private-cloud endpoints with versioned request and response contracts.
CSV, JSON, XML, Parquet, workbooks, batch files and approved result stores can act as governed interfaces when a live service is unnecessary.
For tightly controlled calculations, EVA can create the model request, preserve the approved state, and receive the result after human review and execution.
Every provider is identified by what it is allowed to calculate, what inputs it accepts, what outputs mean, how execution is identified, and how the result can be reproduced or reviewed.
A well-developed workbook may contain years of company, instrument or domain knowledge: revenue bridges, capital structures, scenario logic, financing terms, interactions, caps, floors and analyst judgment encoded in formulas.
EVA can treat that workbook as an external analytical provider rather than forcing a rewrite. The Event layer supplies governed state changes; the spreadsheet performs the calculation; EVA receives the outputs and carries them into the wider consequence analysis.
The model remains familiar. The Event-intelligence layer becomes the additional capability around it.
EVA can sequence several specialist providers without turning them into one monolithic model.
A capacity delay can change operating assumptions, cash flow, financing, valuation and portfolio exposure. Each part can be calculated by the provider already trusted for that task.
EVA owns the Event context, execution sequence, state hand-offs, evidence links, model identities, refusals and final consequence record.
AI can help read documents, identify Events, find evidence, propose mappings and compare precedents. Numerical results remain the output of governed models and analytical services rather than generated text.
Documents, evidence, candidate mechanisms and relevant questions.
Admit the Event, select allowed state changes and choose providers.
Run the approved model with explicit inputs and execution identity.
Carry the result into consequences, stakeholders and decisions.
Kindynos can connect an existing analytical capability to a governed Event workflow without requiring the model to become a Kindynos model.