Donesafe Implementation Recovery: How to Get an Off-Track Project Back Under Control

A project can look busy while still moving away from a workable go-live. Recovery starts by separating what is genuinely broken from what is unfinished, misunderstood or simply outside the original scope.

1. Re-establish the actual scope

Bring the order form, requirements, process maps, decisions and current build together. Identify what was agreed, what has been added, what has been deferred and what remains ambiguous. This stops recovery becoming another round of uncontrolled requests.

2. Review the process before the configuration

If users cannot explain who owns a review, when an action should be created or what closes the process, configuration will not solve the problem. Confirm the future-state process and then map the current build against it.

3. Separate defects from enhancements

A defect is something that does not meet an agreed requirement. An enhancement is a new or changed requirement. Mixing them during UAT makes it impossible to know whether the system is ready. Maintain separate lists with ownership and priority.

4. Find the critical path to go-live

Not every issue has to be solved before launch. Identify the modules, data, integrations, permissions, notifications and reports that are essential for a safe and controlled release. Everything else should be explicitly prioritised into a post-go-live backlog.

5. Rebuild UAT around real scenarios

Test end-to-end workflows with real roles. Include initial submission, review, rejection, escalation, investigation, actions, evidence, close-out, permissions, emails and reporting. A checklist of fields is not enough.

6. Validate data and integrations separately

Where worker, site, department, contractor or training data comes from another platform, prove the interface independently of the front-end workflow. Confirm source of truth, frequency, exception handling, duplicates, terminations and reconciliation.

7. Stabilise reporting before executive sign-off

Reporting often exposes hidden configuration problems. Validate classifications, dates, status logic, action ownership and organisational structures before building management dashboards or promising final metrics.

8. Reset governance

Recovery needs one decision path. Agree who can approve scope changes, who owns requirements, who accepts UAT, who prioritises defects and who has final go-live authority.

What a recovery plan should contain

• Confirmed scope and outstanding requirements
• Current-state configuration assessment
• Defect versus enhancement register
• Critical go-live dependencies
• UAT scenarios and acceptance criteria
• Integration and data reconciliation plan
• Reporting validation
• Decision and escalation model
• Post-go-live enhancement backlog