Data migration
Your history moves with you.
Migration is scoped in discovery, executed in its own phase, and validated before anything cuts over. The working checklist:
Master data
Items, BOMs, routings, work centers, customers, suppliers, and assets are extracted, cleaned, and mapped to the unified data model.
Open transactions
Open sales orders, work orders, POs, and inventory balances load last, so go-live starts from a position that matches reality.
Compliance history
Quality records, CAPAs, training records, and controlled documents migrate to the depth your audit and traceability obligations require.
Validation
Row counts, financial tie-outs, and user spot checks are signed off by your data owners before the parallel run begins.
Integration inventory
Every connection is named before it is built.
Discovery produces an integration inventory: every system that stays, every file feed, every machine or label printer that talks to the platform. Each connection gets an owner, a direction, and a test case, so cutover never discovers a forgotten interface.
CAD, PLM, and estimating systems that feed engineering data
Accounting or corporate ERP staying upstream of the floor
Machine monitoring, scales, and label or barcode hardware
Customer and supplier EDI or portal feeds
Training & AI calibration
Everyone trained, agents tuned to your floor.
Unlimited users means training is not rationed to a license count. Role-specific sessions run during the training phase, while the embedded agents are calibrated against your migrated data.
Role-based training
Operators, supervisors, quality, maintenance, planning, and finance each train on their own workflows - everyone gets a seat.
AI calibration
Scheduling constraints, SPC limits, maintenance thresholds, and inventory policies are tuned to your operation, inside the standard governance pipeline.
Validated cutover
A parallel run against agreed acceptance criteria decides go-live - a sign-off, not a hope.
Deployment architecture
Same method: cloud, on-premises, or air-gapped.
The deployment model changes where the platform runs, not how the implementation is executed. Regulated and export-controlled programs stage the entire toolchain, embedded AI models included, inside their own boundary.
Cloud
Fastest path for operations without data-boundary constraints; managed updates and scaling.
On-premises
The platform and its AI run on your infrastructure, inside your network and access controls.
Air-gapped
No external network calls at all - migration tooling, models, and updates staged inside the boundary for ITAR-controlled and classified-adjacent environments.
Legacy replacement
Replacing a legacy ERP, spreadsheets, or both.
Most deployments replace a patchwork rather than a single system: an aging ERP plus the spreadsheets and point tools that grew around it. The playbook is the same - inventory what each tool actually does, map every job to its module, migrate the data once, and retire the patchwork at cutover instead of running it in the shadows.
The five phases
Adopting Cortrova implementation.
Discovery & planning
Map the operation, name the data owners, build the integration inventory, and lock scope and acceptance criteria.
Data migration
Extract, clean, map, load, and validate - master data first, open transactions last.
Configuration & AI calibration
Workflows, permissions, and compliance settings configured; agents calibrated to your data.
Training
Role-specific training across the whole floor, since users are unlimited.
Validated go-live
Parallel run against acceptance criteria, then a signed-off cutover with all 205+ features live.
FAQ
Questions, answered.
How long does a Cortrova implementation take?
The typical target is 4-8 weeks from kickoff to validated go-live. The window assumes a mid-market operation, engaged data owners on your side, and the standard five-phase parallel deployment. Complex integration inventories or multi-site rollouts are scoped explicitly during discovery, so the timeline you sign is one that reflects your operation.
What happens to our data from the old system?
Data migration is a scoped phase, not an afterthought. Master data - items, BOMs, routings, customers, suppliers, assets, and open orders - is extracted, cleaned, mapped to the Cortrova data model, and validated with row counts and spot checks you sign off on. Historical transactions are migrated to the depth your compliance and costing requirements need.
Do we run both systems in parallel before cutover?
Yes. Cortrova deploys in parallel with your existing systems, and cutover happens only after a validated parallel run: the same orders, inspections, and transactions flow through both, and go-live is signed off against agreed acceptance criteria rather than a calendar date.
What does AI calibration involve?
The embedded agents ship with defaults, then get calibrated to your operation during configuration: scheduling constraints, quality characteristics and SPC limits, maintenance thresholds from your asset list, and inventory policies. Calibration uses your migrated data, and every agent runs inside the same governance pipeline described on our security page.
Can implementation run fully on-premises or air-gapped?
Yes. The same five-phase methodology applies to cloud, on-premises, and air-gapped deployments. For air-gapped installs, migration tooling, model hosting, and updates are staged inside your boundary, so export-controlled technical data never leaves your facility.
Is implementation an extra line item?
No. Implementation is included in the flat annual rate - discovery, migration, configuration, AI calibration, training, and the validated cutover - and all 205+ features are available from day one rather than unlocked over time.
Go deeper
The implementation library.
Manufacturing ERP implementation checklist
The phase-by-phase checklist: what to prepare, who owns what, and where implementations actually slip.
Read →Manufacturing ERP migration guide
How to move off a legacy ERP without losing history, traceability, or the shop floor along the way.
Read →Cloud vs on-premises vs air-gapped ERP
How to pick a deployment model from your data boundary, compliance obligations, and IT reality.
Read →How the five phases run in parallel
The full deployment methodology, phase windows, and the go-live benchmark against the rest of the market.
Read →Security & AI governance
The 7-stage governance pipeline, RBAC, encryption, and audit model your deployment inherits on day one.
Read →Air-gapped deployment
Running the entire platform, embedded AI included, with no external network calls.
Read →Scope it
Tell us your systems. We'll map the path.
Share what you run today, your headcount, and your compliance constraints - we'll scope the phases, the integration inventory, and the timeline to your operation.