Context / Business Challenge
Every enterprise has them: legacy HTML forms scattered across intranets, portals, and SharePoint pages, manually maintained, disconnected from workflows, and far from the self-service experience users expect today. Migrating these forms into a modern platform like ServiceNow has always required significant manual effort — analyzing the form structure, recreating variables, configuring tables, building Record Producers, and setting up Workspaces. Multiply that by dozens or hundreds of forms, and you have a serious bottleneck. The question I wanted to answer was simple: can AI do this for supporting us?
Architecture Blueprint
The solution starts by capturing an HTML form into a centralized Tracker table, ensuring full traceability. An Agentic Workflow is automatically triggered, orchestrating the end-to-end process and invoking the AI Agent. The AI Agent analyzes the form and generates a structured configuration (JSON) representing the target application. The Build Agent executes this configuration to automatically create the Table, Record Producer, and workspace (UX List). A human-in-the-loop validates the outputs, ensuring quality, governance, and alignment before final adoption.
Agentic Workflow
The entry point of the solution is an Agentic Workflow that triggers automatically whenever a new HTML form is registered in a Tracker Table. This table acts as the data contract for the entire process — it captures every form submitted for migration and drives the end-to-end orchestration. Each row represents a migration job, and the status field is updated in real time as the process advances, giving full auditability from the original HTML page all the way through to the final Workspace configuration. No manual handoffs, no human triggers — the workflow fires and the agents take over.
AI Agent
Once the Agentic Workflow kicks in, the AI Agent takes the HTML form as input and performs a deep structural analysis. It identifies every field, its type, its label, and its role within the form — then generates a structured JSON configuration that represents exactly what needs to be built on the ServiceNow side. This JSON becomes the instruction set for the next agent in the chain. What would typically take a developer hours of analysis and documentation is completed in seconds, with consistent output regardless of the complexity of the original form.
Build Agent
The Build Agent receives the JSON and gets to work. It sequentially creates the ServiceNow objects needed to turn that configuration into a live application: the data table, the Record Producer with all the correct variable types mapped from the original HTML fields (using native OOTB ServiceNow variable types), and the Workspace configuration. At every step, the tracker table is updated — status changes in real time, giving a transparent, auditable log of what has been created and when. The result is a fully functional, platform-native self-service experience, built entirely by AI.
Design Principles
- Separation of concerns — each agent has one clear responsibility: the AI Agent analyzes and understands, the Build Agent creates and configures. No agent does both.
- Auditability by design — the tracker table is not just a trigger mechanism, it is the single source of truth for every migration job, capturing state changes end to end.
- Human-in-the-loop by default — AI accelerates the work, but every output is explicitly validated by a human before being considered production-ready.
Key Implementations
- JSON as the contract between agents — the AI Agent stores its structured output directly on the tracker record, decoupling analysis from execution and making the configuration inspectable at any time.
- Build Agent over scripted automation — instead of relying on custom scripts, Build Agent uses native ServiceNow capabilities to create tables, Record Producers and Workspaces, keeping the solution maintainable and upgrade-safe.
Business Impact
Organizations adopting this approach can expect:
- Up to 60% reduction in application development effort
- Up to 70% faster delivery of business solutions
- Up to 50% reduction in configuration and design errors
- Increased productivity of development teams
- Faster response to changing business requirements
Takeaways
The business implications of this kind of automation are significant:
- Effort reduction: What previously required days of manual analysis, development, and testing can now be completed in minutes — with no developer intervention.
- Cost savings: By eliminating repetitive configuration work, organizations can redeploy technical resources toward higher-value activities.
- Scalability: Whether you have 10 forms to migrate or 500, the process runs the same way — consistently, accurately, and without fatigue.
- Auditability: The tracker table provides a complete, step-by-step log of every migration, which is invaluable for governance and change management.
- Faster time to value: Business units get their self-service workflows live faster, reducing the gap between demand and delivery.
- Platform alignment: Every output is built using native ServiceNow capabilities — no custom code, no workarounds, just clean platform configuration that's easy to maintain.
This is AI doing what it does best: taking structured, repeatable human work and automating it at scale.
As I always say in my posts and demos — this is not the only way to do it, and probably not the perfect one. But it's a practical approach that works, and sometimes that's exactly what you need to move fast and deliver real value. 😊
Thank You
Thanks for reading — I hope this demo sparks some ideas about what's possible when you combine ServiceNow's Agentic capabilities with a well-designed automation pattern. If you're exploring similar challenges in your organization, or if you've built something in this space, I'd love to hear from you in the comments. Let's keep pushing the boundaries of what the platform can do. 🚀
Go here to see full demo:
https://www.linkedin.com/pulse/watch-ai-turn-simple-form-complete-business-gaetano-de-mitri-3ahof/
#NowAssist #AI #BuildAgent #AgenticWorkflow #AIAgents
