CEC
Legacy → modern, without losing a single business rule.
CEC (Customer Engagement Center) is a multi-tenant field-service platform — scheduling, dispatch, invoicing, and customer communication — built over years on ASP.NET Web Forms 4.6/4.8. I work on the Web Forms product and on its modernization to a .NET 10 API with a React 19 + TypeScript front end.
- ROLE
- UX modernization · System documentation · React implementation · Workflow redesign
- COMPANY
- Xceleran — field-service software ecosystem
- BEFORE
- ASP.NET Web Forms · .NET Framework 4.6/4.8 · C# · SQL Server · jQuery · Bootstrap
- MODERNIZING TO
- .NET 10 API · React 19 · TypeScript · Vite · TanStack Router & Table · React Hook Form + Zod · Tailwind CSS
- STATUS
- Ongoing — starting with scheduling and appointments
AT A GLANCE
~117
ASPX pages
in the legacy web application
100+
Database tables
across three SQL Server databases
7
External integrations
accounting, SMS, email, meetings, payments, financing
30+
Mobile API endpoints
serving the field technician app
FIGURES FROM THE SYSTEM AUDIT — NO CUSTOMER DATA SHOWN
The old product was working.
That was the problem.
Field-service companies run their day on CEC: booking jobs, dispatching technicians, sending reminders, invoicing, and syncing to accounting. Years of real business rules live inside its pages, code-behind files, and stored procedures.
The question wasn’t whether to modernize — full-page postbacks, fixed layouts, and scattered logic were slowing everyone down. It was how to do it without throwing away the knowledge embedded in a system people depend on every day.
Much more than a calendar
Before redesigning anything, every capability was inventoried. This is the surface a modern UI has to carry.
Scheduling & dispatch
- Drag-and-drop calendar (FullCalendar Scheduler)
- Real-time dispatch board
- Technician availability and time blocks
- Auto-generated Zoom / Google Meet links
Customer CRM
- Residential customers and B2B contacts
- SMS, email, and chat history on the profile
- Bulk customer import
- File and photo attachments
Quotes, invoices & payments
- HTML-to-PDF estimates and proposals
- Job-to-invoice conversion
- Card payments and consumer financing
- Two-way QuickBooks Online sync
Automated communication
- SMS and email appointment reminders
- Post-job follow-ups
- Customer satisfaction surveys
- Inbound SMS pushed live to dispatchers
Operations & reporting
- Dynamic report builder
- Purchase orders
- Items, groups, and services
- Sales and technician commissions
Field technician app
- Schedule sync
- Job photos
- On-site invoicing
- Payment links
Five projects, one monolith
Understanding where logic lived — page, processor class, background job, or database — decided what could move to React and what had to stay behind an API.
01 FRONTEND
- ASP.NET Web Forms
- jQuery · Bootstrap
~117 pages, with rules split between processor classes and code-behind.
02 API
- ASMX mobile API
- SignalR hub
30+ JSON endpoints for the technician app; real-time pushes to dispatch.
03 DATA
- SQL Server × 3
- 100+ tables
Scheduling, jobs, and core data behind a custom ADO.NET wrapper.
04 INTEGRATIONS
- 7 external services
Accounting, SMS, email, meetings, payments, and financing.
05 JOBS
- 3 console apps
Scheduled reminders and notification emails, audit-logged against duplicates.
ONE REQUEST, END TO END
- Customer replies by SMS
- SMS integration
- SignalR broadcast
- Every open dispatch board updates
MAIN WEB APP
ASP.NET Web Forms
~117 pages; business logic split between processor classes and page code-behind; menus gated by 10 session privilege flags.
MOBILE API
ASMX web services
30+ JSON endpoints for technician login, schedules, photos, invoices, and payment links.
BACKGROUND JOBS
3 console apps
Scheduled appointment reminders and notification emails, with audit logs to prevent duplicates.
REAL-TIME
SignalR
Inbound customer SMS broadcast to every open dispatch board.
ADMIN
Classic ASP module
System-level tenant administration kept from an earlier generation.
DATA
SQL Server × 3
Scheduling, jobs, and core data in separate databases, accessed through a custom ADO.NET wrapper.
Business rules must be preserved
Every validation, conditional, and workflow state has to behave the same way in the new UI.
No workflow regressions
Dispatchers built their day around specific sequences. The new version supports them rather than reinventing them.
Incremental, not big-bang
The business keeps running on the legacy system while modules move over one at a time.
Improve within the rules
Hierarchy, feedback, and responsiveness improve freely — business logic does not change silently.
- 01
Write the system down first
Nine reference documents before any redesign: architecture, business rules, database, features, glossary, UI inventory, workflows, tech stack, and open questions.
- 02
Separate accidental from intentional complexity
Some complexity is the business. Some is years of workarounds. Each rule was traced to its source before deciding what to keep.
- 03
Rebuild the core workflow first
Scheduling is the heart of the product, so appointments — dashboard, list, details, booking wizard, settings — moved first.
- 04
Verify rules against both systems
If a rule produced a result in Web Forms, the React screen must produce the same result. Paths are checked side by side.
- 05
Improve where the rules allow
Where logic does not dictate presentation — hierarchy, spacing, feedback, responsiveness — the new UI is free to be better.
THE CENTRAL DECISION
“The business rules are the product. Web Forms is just the container. Replace the container.”
Same jobs, less friction
Each workflow keeps its rules and its order of operations. What changes is how much the interface makes people wait, scroll, and remember.
Booking an appointment
BEFORE · WEB FORMS
One dense page: choose customer, service type (which sets duration), and technician. Changes trigger full-page postbacks before availability is known.
AFTER · REACT
A multi-step Appointment Wizard — customer, service, technician and time — validated inline with React Hook Form + Zod, with no page reloads between steps.
Finding appointments
BEFORE · WEB FORMS
Server-rendered grids with postback paging; filters reset the page.
AFTER · REACT
A sortable, filterable TanStack Table list, with a dedicated details view for each appointment.
Booking rules & settings
BEFORE · WEB FORMS
Service types, time blocks, and statuses live on separate admin pages.
AFTER · REACT
One Appointment Settings screen that groups the rules that shape booking.
Navigation & access
BEFORE · WEB FORMS
A master page shows or hides menu items from session privilege flags.
AFTER · REACT
Typed routing with TanStack Router and a dashboard entry point; the documented privilege model defines what each role can reach.
Layout
BEFORE · WEB FORMS
Fixed-width pages built for one desktop resolution.
AFTER · REACT
Responsive Tailwind layouts with a shared component set and Framer Motion feedback.
Analysis, documentation, redesign, and implementation
System analysis
Traced workflows, rules, and data dependencies across pages, processors, background jobs, and the mobile API.
Documentation
Turned the audit into reference docs the whole team can build against — features, rules, workflows, glossary.
UX modernization
Redesigned hierarchy, navigation, and interaction models without changing what the business depends on.
React implementation
Building the new frontend in React 19 + TypeScript on Vite, with TanStack Router, TanStack Table, and schema-validated forms.
Workflow redesign
Multi-page sequences restructured into single flows with client-side state — fewer steps where possible, preserved where required.
Rules documented end to end.
The core workflow rebuilt first.
The React foundation — routing, tables, forms, motion, and a shared component set — is in place, and the appointments module leads the migration while the legacy system keeps the business running.
Understand before you redesign.
Redesigning a legacy system without mapping it is the fastest way to break something that was working.
Technical and design debt compound together.
The interface problems reflected structural problems in the code. Fixing one without the other only goes so far.
Migration is an opportunity for clarity.
Moving to React forced every rule to be written down — surfacing assumptions that had never been documented.