10decodersOne template ID, identical in every environment
A leading U.S. supplier of home-delivered urologic and ostomy products runs a document intelligence platform that depends on templates behaving the same way from Development to Production. 10Decoders built a synchronisation engine that keeps template IDs and schemas aligned across all four environments automatically, without ever slowing down the people using them.
Same template, same number, everywhere
The client's platform reads medical forms with AI: extraction models, text-region detection and signature checks all follow structured templates. Every template has an ID, and every downstream step trusts that ID to point at the right form. When a template built in testing got a different ID once it reached Production, results landed against the wrong schema and audit trails broke.
10Decoders made Production the single source of truth for template IDs, then built the plumbing to reserve, distribute and secure those IDs across Development, UAT, Staging and Production — all in the background, with a safe fallback when a remote environment is unreachable.
Six pieces that keep environments in step
Production as Canonical Registry
Production allocates every global template ID, so no two environments ever hand out the same number twice.
Canonical ID Reservation
Lower environments request the next global ID before creating a template, then align their own sequences to match.
Skeleton Reference Copies
Production pushes an inactive placeholder with the canonical ID and metadata to every other environment.
Degraded-Mode Fallback
If Production can't be reached, template creation falls back to local allocation and finishes in milliseconds.
Dedicated Service Accounts
Each environment provisions its own service account on first sync, so no user credentials cross environment lines.
Signed Token Checks
Every inter-environment request is cryptographically signed and independently verified before any data is accepted.
Template ID 45 meant one form in UAT and a different form in Production.
The Challenge
- Each environment's database handed out its own template IDs, so the same template drifted to different numbers.
- Extraction pipelines referencing a drifted ID mapped medical data to the wrong form schema.
- Moving templates between environments meant manual scripts, environment freezes and release delays.
- Synchronous cross-environment calls risked stalling local users whenever a remote environment was slow or down.
Our Approach
- Made Production the only place global template IDs are issued, with every environment reserving from it.
- Guaranteed every environment uses the same canonical ID for the same template, so schemas always match.
- Replaced manual migration with automatic skeleton distribution running in the background.
- Kept every sync step non-blocking, with a local fallback so template creation never waits on the network.
The Solution
Before a template is created in any lower environment, the platform sends a short, authenticated request to Production to reserve the next canonical ID. Production records the reservation and returns the number; the originating environment saves its template under that exact ID and moves its own database sequence forward so it never collides later.
Once a template exists, Production broadcasts an inactive skeleton copy to every other environment, holding that ID so nobody can overwrite the slot by accident. Teams configure and test templates the way they always have — the alignment happens on its own, and every lifecycle event is logged with a trace that follows it across environments.
Reserve
Lower environment asks Production for the next global template ID.
Create & Align
Template is saved under the canonical ID; the local sequence is moved past it.
Distribute
Production sends an inactive skeleton copy to every other environment.
Verify & Trace
Signed requests are checked on arrival; every event is logged end to end.
From local edit to global alignment
Dev · UAT · Staging
Teams create and test templates locally, requesting a canonical ID first.
Production Canonical Registry
Reserves IDs, verifies signed requests and broadcasts skeleton copies.
Aligned Schemas Everywhere
Extraction pipelines in every environment resolve each ID to the same form.
Built to be correct and fast at once
One Issuer, No Collisions
A single canonical registry removes the root cause of ID drift instead of patching it after the fact.
Nothing Waits on the Network
Sync runs asynchronously, so local API responses stay sub-second regardless of remote latency.
Graceful When Production Is Down
Degraded mode lets work continue locally instead of throwing errors during outages or maintenance.
Reserved Slots Can't Be Overwritten
Skeleton copies hold each canonical ID in every database before anyone can reuse it.
Zero-Trust Between Environments
Dedicated service accounts and signed tokens mean each environment verifies callers for itself.
Every Change Is Traceable
Global trace tracking follows each template lifecycle event across all environments for audits.
Three capabilities, one engine
Canonical ID Reservation
Production-issued global template IDs with automatic sequence alignment in every lower environment.
Asynchronous Skeleton Distribution
Background broadcast of inactive reference copies that reserve each ID across all databases.
Zero-Trust Service Security
Auto-provisioned service accounts and cryptographic token checks on every cross-environment request.
What changed for the platform
Zero ID collisions
Every environment now resolves a template ID to the same form, with one source of truth behind it.
Zero processing delays
Local workflows stay sub-second because synchronisation never sits in the request path.
Four environments aligned
Development, UAT, Staging and Production stay in continuous step without coordination meetings.
No more manual scripts
Template distribution is fully automated — no hand-run database updates or environment freezes.
Extraction data stays correct
AI pipelines no longer map medical form data to the wrong schema between UAT and Production.
Audits get simpler
Each template's full lifecycle is logged with a trace that spans every environment.
How it was delivered
Delivery Quality
- Live in production, synchronising real medical templates in the client's infrastructure.
- Rolled out without blocking day-to-day template configuration or testing.
- Built on top of the client's existing document intelligence platform.
Security & Compliance
- No user credentials ever travel between environments.
- Every inbound sync request is signed and independently verified.
- Lifecycle events carry a global trace for regulatory review.
IP & Reusability
- Designed as a reusable core platform capability.
- Now a standard 10Decoders pattern for multi-environment AI schema management.
- Roadmap: field-level diffing and one-click versioned rollback.
Value for every team that touches templates
Configure freely, ship sooner
Teams build and test templates anywhere without scheduling migrations or freezing environments.
Extraction you can trust
Medical documents are always read against the right form, so downstream workflows fire correctly.
A clear trail for audits
Every template change is traceable across environments, simplifying regulatory review.
