Controlled migration runbook

Switch gambling compliance platforms without losing the compliance record.

Preserve the compliance record, prove priority workflows, control the dual run and keep a reversible path until the new platform is accepted.

The migration is complete only when records reconcile, intended users complete the live work and historical decisions remain retrievable—not when a bulk import reports success.

Public product information reviewed

Six-phase migration

Every phase ends with evidence and an acceptance gate.

Do not move from export to cutover in one jump. Prove the data model, priority workflows and rollback path in bounded stages.

Inventory the compliance record

Count records, attachments, history, users, integrations and reports. Identify legal, audit and retention constraints.

Evidence: Signed inventory, sample exports and named owners for every data class.

Map old to new

Map identifiers, statuses, taxonomies, permissions, relationships and workflow states before transforming data.

Evidence: Approved field map with explicit decisions for every unmapped value.

Migrate a representative slice

Use difficult and historical records, not only clean recent examples. Reconcile counts and field-level results.

Evidence: Priority samples are complete, readable, linked and independently checked.

Run the operating scenarios

Repeat research, change, assignment, evidence, sign-off, reporting and export in the new platform.

Evidence: Named users complete the agreed cases against written success criteria.

Control the cutover

Freeze or reconcile changes, define the authoritative system, communicate ownership and preserve rollback.

Evidence: Signed go/no-go, working integrations, support coverage and tested rollback steps.

Close and retain

Confirm exports, archive access, record retention, access removal, deletion evidence and contract closure.

Evidence: Legacy system receives no live work and the retained archive passes retrieval tests.

The compliance record

Move relationships and history—not isolated rows.

Markets, authorities, licences, activities and products

Requirements, governing material, versions and checked dates

Applicability decisions, assumptions and unresolved questions

Tasks, owners, deadlines, comments, approvals and status history

Evidence, attachments, reports and audit events

Users, roles, permissions, integrations and retained identifiers

Outcome evidence

Migration success needs a reconciled acceptance receipt.

Like-for-like buyer outcome · proof pending

No public benchmark predicts migration time or success for an unknown source system and data set. Measure record reconciliation, field completeness, workflow completion, historical retrieval, user acceptance, cutover incidents and remaining manual work.

Retain the inputs, elapsed time, corrections, manual work and completed output from every shortlisted supplier. Marketing claims do not replace a buyer-owned test.

Buyer-owned validation

Keep unknowns visible until the evidence arrives.

A supplier can answer privately, in a contract or during a live test. Until then, the correct state is unknown—not zero and not an inferred product gap.

Use identical scope, inputs and preparation time.

Record what exists now, what needs services and what is only planned.

Keep product proof separate from implementation promises.

Contract data return, renewal and exit before selection.

Evidence and disclosure

Claims need proof.

Product pages describe scope. The buyer’s own case must prove accuracy, usability and fit.

Atlas publishes this analysis and is not affiliated with, sponsored by or endorsed by the named vendors. No vendor paid for placement. Customer reports are identified as reports; unknown information remains unknown. Corrections follow the corrections policy.

Questions buyers ask

Key questions, answered.

How long does it take to switch compliance platforms?

The duration depends on data volume, integrations, markets, workflow complexity and vendor support. Build a phased plan with inventory, mapping, migration tests, a controlled dual run, acceptance criteria and a reversible cutover rather than relying on a generic timeline.

What data should be exported before switching?

Export requirements, sources, versions, market and licence mappings, decisions, tasks, owners, dates, comments, approvals, attachments, evidence, users, permissions, reports and audit history, together with an explanation of every field and identifier.

Should old and new compliance systems run in parallel?

A bounded dual run is usually safer for material workflows. Define which system is authoritative, prevent duplicate actions, compare outputs on known scenarios and set an end date so parallel operation does not become permanent.

How can a team avoid losing audit history?

Preserve immutable exports, identifiers, timestamps, actors, approval states, attachments and source references. Test retrieval of selected historical decisions before accepting the migration and retain the legacy archive under a documented policy.

What is a safe compliance-platform cutover?

A safe cutover has signed acceptance criteria, reconciled record counts, tested priority workflows, trained owners, working integrations, rollback steps, a support window and explicit confirmation that the old system is no longer receiving live work.

Run the operating case

Put Atlas through the same scored test.

Bring one difficult market, one material change and the outputs your team needs to retain. We will show the completed workflow and identify what remains outside Atlas.

Test the case in Atlas