Background
Four business units, each running their own processes on paper, email, and spreadsheets. HR, Finance, Operations, and Procurement — each with their own workflows, their own filing systems, and their own informal way of chasing approvals when things got stuck.
Delays in one unit cascaded to others. Approvals disappeared into inboxes. Management had no real-time view of what was in progress, what was blocked, or what had actually been completed. The audit trail for any given process was whatever someone could piece together from email threads.
Anil Choudhary led the discovery, design, and delivery of the automation platform across all four business units.
The Challenge
Manual Processes at Scale
The 15 workflows identified for automation covered a significant volume of work:
| Business Unit | Key Workflows | Avg. Processing Time |
|---|---|---|
| HR | Onboarding requests, leave approvals, equipment requests | 3–5 days |
| Finance | Purchase approvals, expense submissions, vendor onboarding | 4–7 days |
| Operations | Incident reports, maintenance requests, shift change requests | 2–4 days |
| Procurement | RFQ creation, supplier approval, contract renewals | 5–10 days |
Each workflow involved multiple approvers, document attachments, and status tracking — all done via email and spreadsheets.
Visibility and Accountability Gaps
Management had no reliable view of operational throughput:
- No dashboard existed showing requests in progress vs. completed vs. blocked
- Bottleneck identification required manually reviewing email threads
- SLA compliance was untracked — there was no automated escalation when requests aged past acceptable thresholds
- Audit evidence for completed workflows was assembled manually from email archives before each review cycle
Data Quality and Rework
Manual data entry across disconnected systems created constant rework:
- Forms were filled out differently by different teams — inconsistent field formats, missing information
- The same data was entered multiple times as it passed between systems
- Errors were discovered downstream, requiring processes to restart from earlier stages
- No validation logic prevented obviously incorrect entries from entering the workflow
Solution Architecture
The Power Platform solution was designed in three layers: digitized input (PowerApps), automated orchestration (Power Automate), and operational visibility (Power BI), all backed by SharePoint as the document and data store.
Data and Document Foundation: SharePoint
Before building any apps or flows, a structured SharePoint environment was established:
- Dedicated SharePoint site per business unit with consistent list schema
- Shared document libraries with version control for attachments
- Managed metadata taxonomy applied across all lists for consistent categorization
- Permission model: requestors can submit and view their own items; approvers see their queue; managers see all
This provided a clean, governed data layer that all downstream automation could rely on.
Digitized Forms: PowerApps
Canvas apps replaced paper and email-based form submission for all 15 workflows. Key design principles:
- Validation at entry — required fields enforced, format validation on dates and amounts, lookup fields preventing free-text errors
- Context-aware forms — fields show or hide based on prior answers (e.g., a purchase request for over £10k automatically adds a CFO approval stage)
- Offline capability — field operations staff can complete forms without connectivity; data syncs on reconnection
- Mobile-first layout — all apps designed for phone and tablet use, tested on iOS and Android
PowerApps Canvas App Structure (example: Purchase Approval)
├── Screen 1: Request Details
│ ├── Department (dropdown — lookup)
│ ├── Vendor (lookup from Approved Vendor list)
│ ├── Amount (number field with currency format)
│ └── Justification (multi-line text, required)
├── Screen 2: Supporting Documents
│ └── Attachment control → SharePoint document library
├── Screen 3: Review & Submit
│ └── Summary view → triggers Power Automate flow on submit
└── My Requests screen
└── Gallery showing user's submissions with status
Workflow Orchestration: Power Automate
Each submitted form triggered a Power Automate cloud flow handling routing, approvals, notifications, and data updates automatically:
Purchase Approval Flow
Trigger: PowerApps form submitted
│
├── Get approver from department mapping list
│
├── [If amount > £10,000]
│ └── Add CFO as secondary approver
│
├── Send approval request (Adaptive Card in Teams)
│ ├── Approved → update SharePoint status, notify requestor, continue to next approver
│ └── Rejected → update status, notify requestor with rejection reason, end flow
│
├── [All approvers approved]
│ ├── Update SharePoint item status = "Approved"
│ ├── Notify requestor and Finance team
│ └── Create task in Planner for Finance processing
│
└── [SLA timer — 48 hours]
└── Escalate to approver's manager if no action taken
SLA escalation was one of the highest-value automations — previously, stalled approvals were noticed only when the requestor followed up manually. Automated escalation eliminated this entirely.
Notification and Approval Experience: Teams Integration
Approvals were delivered directly in Microsoft Teams via Adaptive Cards, so approvers could approve or reject without leaving the platform they were already in:
- Request summary displayed inline in the card
- Supporting documents linked (not attached — no inbox bloat)
- Approve / Reject buttons with a required comment on rejection
- Mobile-compatible — approvals actionable from the Teams mobile app
Operational Dashboards: Power BI
A shared Power BI workspace provided real-time operational visibility across all four business units:
Operations Overview Dashboard
- Total requests by status (pending / approved / rejected / overdue)
- Average processing time by workflow type — trend over 30/60/90 days
- SLA breach rate by department
- Top bottlenecks: which approval stages had the highest average wait time
Individual Unit Dashboards Each business unit had its own page with granular metrics relevant to their workflows — Finance tracked approval volumes and spend by category; HR tracked onboarding completion times; Procurement tracked vendor onboarding cycle time.
Data refreshed every 30 minutes from SharePoint via the Power BI connector.
Workflow Inventory
| # | Business Unit | Workflow | Automation Type |
|---|---|---|---|
| 1 | HR | New employee onboarding request | Multi-stage approval + task generation |
| 2 | HR | Leave and absence request | Line manager approval + calendar integration |
| 3 | HR | IT equipment request | IT queue + auto-assignment |
| 4 | Finance | Purchase order approval | Multi-tier approval based on amount |
| 5 | Finance | Expense claim submission | Manager approval + Finance processing |
| 6 | Finance | Vendor onboarding | 4-stage cross-department approval |
| 7 | Finance | Contract renewal alert | Automated trigger 60/30/7 days before expiry |
| 8 | Operations | Incident report | Automatic routing by severity |
| 9 | Operations | Maintenance request | Team assignment + progress tracking |
| 10 | Operations | Shift change request | Supervisor approval + schedule update |
| 11 | Procurement | RFQ creation and distribution | Vendor notification + response collection |
| 12 | Procurement | Supplier performance review | Scheduled trigger + form distribution |
| 13 | Procurement | Contract document management | Approval + SharePoint filing + expiry tracking |
| 14 | Cross-unit | External visitor management | Security notification + badge generation |
| 15 | Cross-unit | Policy acknowledgement tracking | Bulk send + completion tracking |
Results
| Dimension | Before | After |
|---|---|---|
| Average process handling time | 3–7 days | Under 24 hours (routine), under 4 hours (urgent) |
| Reduction in handling time | — | 70% average |
| Paper-based processes | 15 workflows | Zero |
| Manual data re-entry | Every workflow | Eliminated — data entered once |
| Approval visibility | None — email inbox only | Real-time Teams notifications + Power BI dashboard |
| SLA escalation | Manual follow-up | Automated at 48-hour threshold |
| Audit trail completeness | Partial — email archive | Full — every action timestamped in SharePoint |
| Rework due to data errors | Frequent | Rare — validation at point of entry |
The most visible change was the elimination of paper. Forms that used to require printing, signing, scanning, and emailing were replaced by mobile apps that could be completed and approved in minutes. But the Power BI dashboards ended up being equally impactful — they surfaced bottlenecks that management had always suspected existed but couldn't quantify. Once visible, those bottlenecks became fixable.
