Eight named deliverables. Start with a demo rebuild or a tenant extract — everything after builds on the same repository.
Source books → populated Acumatica test tenant the prospect can click through, then a path to go-live.
Test/sandbox write · not production
Details →Business scenarios run on an upgraded copy before each major — pass/fail per tenant for your release gate or install base.
Runs on upgraded copies, never production
Details →Full tenant config extracted into Git, plus a findings report — custom, drifted, undocumented.
Read-mostly access
Details →Cutover rehearsed until boring: rebuild, load opening balances, controller-signable reconciliation.
Scoped per engagement
Details →Industry demos rebuilt on demand — empty to realistic, linked transactions. Optional proof tenant from the prospect’s own books.
No production access
Details →Weekly read-only sandbox vs production diff — material drift ranked above noise.
Read-only
Details →Scenario suites for your extension on each release candidate — wired into the pipeline you already run.
Plugs into your pipeline
Details →Green/red checks for one named integration pair first — Shopify, EDI, WMS, CRM, payments — on every release.
Scoped credentials per integration
Details →Nothing installs inside Acumatica. Nothing for you to keep running.
Work product stays in your repo. Read-only work only reads; release and cutover never touch production until the SOW says so. Hosted tenants use what you already have; full rebuilds need a lab you control — every engagement states which — and your Acumatica version is pinned.
Same four building blocks under every offer above — not separate products you buy.
QuickBooks or Xero field mapping sits inside the cutover loop when the engagement needs it. Rebuild the tenant, load what was scoped, exercise it, prove the books, tear down — then do it again until go-live is boring.
I rehearse cutovers on rebuildable hosts with seed data and recorded scenarios.
Live tenant out to Git, or a clean tenant up from versioned config. Every change gets the same linked transactions before it is applied — so drift and release work share one tree.
I keep tenant configuration as versioned YAML and build, test, and apply from that tree.
Provision and rebuild the hosts the offers run on — cloud or on-prem — with two-layer backups and an off-site copy. Full rebuilds need a lab you control; every engagement states which path.
I automate provisioning, virtualization, and tested off-site backups for Acumatica hosts.
Tool access, retrieval, orchestration, and guardrails for a single business path — scenario generation and field mapping in the offers above, or a non-ERP workflow when that is the job.
I built MailPilot end-to-end: inbound mail answered from an indexed knowledge base in under 60 seconds (measured), declined when the answer is not in the data.