The CreatorCon Call for Content is officially open! Get started here.

What do you do with more than 500 legacy ServiceNow workflows?

kevinanderson
Giga Guru

Organizations that have been using ServiceNow for ten years or more now have often accumulated hundreds of legacy workflows that, over time, are becoming more difficult to maintain.  Flow Designer has become the modern standard for building and automating workflows, but large orgs may have many business processes still relying on the legacy workflow technology.  To stay current and take advantage of new platform capabilities, it's essential to begin aligning these legacy workflows with Flow Designer.

 

Our modernization process begins with a thorough analysis of each legacy workflow to document its logic and requirements.  A key innovation in our modernization approach is a two-phase workflow analysis powered by multi-modal generative AI. In the first phase, we use multi-modal generative AI to process both the workflow screenshot and its XML extract, automatically converting these technical artifacts into a clear, text-based representation of the underlying workflow logic. This “workflow-to-text” format makes it much easier for the AI to understand how the legacy processes operate.  We then take this structured workflow-to-text document and run it through generative AI again, translating the technical logic into a concise business process summary. This higher-level workflow summary helps enable more rapid flow buildout, accelerate the authoring of QA test case documentation, and help make it easier for stakeholders to review for final signoff.

 

Once the business logic has been captured, we leverage a set of prebuilt flow templates—including request, approval, fulfillment, and confirmation flows—to accelerate development. By using a "mock-up" approach, we translate the business logic into the new flows (copied form our templates) with placeholder steps, placeholder data pills, and a pseudo-code-like frame-out. This creates a quick skeleton in Flow Designer, allowing us to rapidly migrate the core logic structure of the legacy workflow into flow designer.  

 

Another essential part of the migration process is the migration of the legacy workflow scripts into new flow-action script containers within Flow Designer. These legacy scripts do a large variety of things, including if-else logic, assignment groups, request and task data updates, integrations, etc.  Maintaining this script logic without forcing a re-write is critical for a quick migration. Using a prebuilt template, we leverage generative AI to assist converting the existing workflow scripts into flow action scripts, so they continue to work as build in original legacy workflows but now those scripts return data pills for consumption by flow designer actions and sub flows. 

 

With the new flows built, rigorous testing is essential to ensure they faithfully reproduce the behavior of the legacy workflows. To support this, we use a custom data mining script to extract representative ticket data from a recent production clone.  Using the workflow analysis document, the form variables names and variable options, and a template script build for capturing existing requests as a JSON blob, we employ Generative AI to author a custom data mining script that script gathers real-world request records, scrubs any sensitive information, and outputs a set of test data covering all logic paths in the workflow. 

 

We then use this data to quickly auto-generate test tickets in the development environment, allowing us to compare the new Flow Designer implementation against the original legacy workflow example tickets in production (or a prod-clone). By validating approvals, task assignments, and process outcomes with real data, we can quickly identify any discrepancies and refine the new flows until they match legacy behavior

 

After the new flow has been built and initial testing is complete, the next step is to create a comprehensive QA test case document.  We use generative AI in combination with our workflow analysis document, and a QA test case template to quickly generate a complete list of test cases for the target workflow. This document is essential part of the handoff to QA team to ensure testing is conducted quickly while guaranteeing that all logic pathways are covered in the regression testing done by our QA team members.  Handoff to the QA team includes the analysis document, QA test case document, and the list of sample tickets from production that were used for unit testing.

 

At the heart of our modernization strategy is the transformative use of AI to accelerate migration and maximize throughput. By automating analysis, documentation, and script conversion, we can process and migrate a large volume of workflows efficiently.  This AI-driven approach allows us to deliver rapid results, maintain quality, and ensure that our workflow modernization effort can be completed in a timeframe that is hopefully a fraction of the time this migration would have taken in the era before the advent of generative AI. 

 

AI tooling for this project:  Open AI Chat GPT 4.1, multi-modal web UI

 

Legacy Workflow Modernization Quick Checklist

  1. Workflow Analysis

    • Capture screenshot and XML extract of legacy workflow
    • Use multi-modal AI to convert to workflow-to-text format
    • Generate business process summary via generative AI
  2. Template Setup

    • Copy request, approval, fulfillment, and confirmation flow templates
    • Build skeleton in Flow Designer using placeholders and pseudo-code
  3. Script Migration

    • Migrate legacy workflow scripts to flow-action containers
    • Use AI to convert script logic and ensure standard logging
  4. Flow Buildout
    • Migrate steps in the legacy workflow to the approvals and fulfillment sub-flows
  5. Test Data Preparation

    • Author data mining script using workflow analysis and form details
    • Extract and scrub production ticket data for all logic paths
  6. Unit Testing

    • Auto-generate test tickets in development environment
    • Compare new flows against legacy workflow behavior and refine as needed
  7. QA Documentation & Handoff

    • Assemble analysis, test case, and sample ticket documents
    • Deliver QA package and coordinate regression testing
  8. Defect Resolution & UAT

    • Address QA findings and retest as needed
    • Obtain signoff from business/process owner for production cutover

By combining automation, AI, flow designer templates, and a muti-step workflow analysis and build out process, this modernization initiative helps ensure legacy workflows are transformed quickly and reliably.  By migrating these legacy workflows to Flow Designer, we aim to facilitate greater maintainability, faster adaptation to business changes, and help our development team focus on delivering value though modern tooling.

 

 

0 REPLIES 0