co-existence and integrations
mojoh defines types and flows. your data and model layers stay where they are. this page explains how mojoh fits in.
Most teams first apply these patterns to ERP migrations and integrations while keeping their existing warehouse and iPaaS in place.
with warehouses and lakehouses
mojoh defines types and flows; warehouses and lakehouses remain your systems of record for stored data and analytics. flows read and write through typed endpoints. mojoh tracks lineage and rollback at the flow level.
Instead of doing data quality in spreadsheets, define validation and cleansing steps here and run them against warehouse tables as part of mojoh flows.
common patterns:
- Use idempotent upserts into warehouse/lakehouse tables from mojoh flows.
- Use typed snapshots to support rollback and comparison between migration runs.
- Define data quality checks (validation rules, failed-record capture, reporting) on top of models and flows.
- Apply reusable cleansing and enrichment steps before and during migrations.
with model platforms
model platforms (for example Bedrock or Azure OpenAI) handle inference. mojoh calls them from flows—not the other way around. prompts, inputs and outputs are logged against motes so you can audit and replay model interactions.
common patterns:
- Call Bedrock/Azure OpenAI models from mojoh flows to enrich or classify records, then write results back as typed motes.
- Keep prompts, inputs and outputs logged against motes and links so you can audit and replay model interactions.
- Anchor guardrails to mojoh's types for consistent validation.
with iPaaS
keep edge adapters and SaaS connectors in your existing iPaaS. use mojoh flows for typed transformations, data quality checks and lineage, then hand results back to iPaaS for delivery.
common patterns:
- Let your existing iPaaS handle edge adapters and SaaS connectors.
- Use mojoh flows for typed transformations, data quality checks and lineage, then hand results back to iPaaS for delivery.
- Trigger mojoh flows from iPaaS when ERP or SaaS events occur, so transformations run on governed models instead of point-to-point scripts.
reference patterns (examples)
Append-only migration from legacy ERP to D365 with typed snapshots for rollback
Typed transformations through your data warehouse with full lineage and rollback.
Data quality checks on warehouse tables with failed-record capture in mojoh
Define validation rules here and run them instead of doing it in Excel.
Enrichment of customer records via model platforms, logged against motes
Model inference integrated with mojoh flows, types and audit trails.
iPaaS at the edge, mojoh at the core
Keep edge adapters in your iPaaS while centralizing typed transformations in mojoh.