Case Studies / Internal Operations Platform
MSP Operations & Billing Automation Platform
A custom internal platform for a ~100-person MSP that transformed a week-long manual billing reconciliation into a controlled, auditable, automated process.
The Situation
A managed service provider with approximately 100 employees manages Microsoft 365 licensing across a large number of customer accounts. Each month, licensing quantities must be reconciled between the distributor billing system, the PSA platform, and the actual Microsoft tenant data.
This process is a core financial control point. Errors result in customers being billed for licenses they don't have, or licenses going unbilled. Either outcome creates friction and erodes trust.
The Problem
Before the platform was built, the reconciliation was entirely manual. A staff member would export data from the distributor system and the PSA platform, load both into spreadsheets, and compare quantities row by row across hundreds of line items.
The process regularly took the better part of a week. Discrepancies had to be investigated manually. Updates had to be entered manually into the PSA. The entire process was undocumented, non-repeatable, and dependent on the individual performing it.
Billing errors were a regular occurrence. Some were caught before invoices went out; others were not. When errors were discovered after invoicing, the resulting corrections required customer communication that created reputational risk.
The Solution
A custom internal platform was designed and built specifically for this organization. No off-the-shelf product adapted to fit. No generic integration tool configured around someone else's assumptions. The platform integrates distributor billing data, PSA agreement data, and Microsoft tenant data into a unified reconciliation workflow built around the exact way this business operates.
The platform automatically pulls current quantities from each source, applies the business rules for how discrepancies are classified, and produces a structured reconciliation report. A staff member then reviews the report, approves or adjusts proposed changes, and triggers the controlled production run.
The human role shifted from manually comparing spreadsheets to auditing a structured output and making judgment calls on flagged exceptions. The tedious, error-prone part was automated. The decision-making and oversight remained human.
Control Architecture
Every reconciliation run follows a staged execution model. No data is written to any live system without a human review and explicit approval step.
"Automate aggressively. Keep consequential actions controlled."
This pattern (dry run, validate, approve, execute) applies to most production write operations in the platform. It is not optional overhead. It is the mechanism that makes automation trustworthy in a financial context.
The Billing Cycle interface guides operators through each stage in order. Steps 1–4 are complete. Step 5 (Publish) is unlocked and ready — all previews passed with zero errors.
Result
The same reconciliation that previously required up to a week of manual spreadsheet work now completes in a structured two-phase process.
The previous duration is not presented as a precise before/after comparison because the exact prior labor hours were not formally tracked. What is clear is that the process shifted from a multi-day manual exercise to a controlled workflow that a single staff member can complete in a structured morning.
Platform Capabilities
Billing reconciliation was the highest-priority workflow, but the platform was designed as a full internal operations system. It is currently in limited production use and preparing for organization-wide deployment to approximately 100 users.
Technician Utilization View
| Engineer | Billable hrs | Non-billable | Total | Utilization |
|---|---|---|---|---|
| A. Rivera | 72.5 | 8.0 | 80.5 | 90% |
| B. Chen | 68.0 | 12.5 | 80.5 | 84% |
| C. Torres | 54.0 | 18.0 | 72.0 | 75% |
| D. Kumar | 48.5 | 22.0 | 70.5 | 69% |
Utilization data is pulled automatically from PSA time entries. Engineers and managers see role-appropriate views of the same underlying data.
Financial Insights View
Financial figures update automatically each month as billing cycles close. All data is sourced from approved billing snapshots, not live system estimates.
Built for this organization. The same approach applies to yours.
Every capability here was designed around this organization's specific workflows, data sources, and operational requirements. The reconciliation logic reflects their specific distributor relationship and PSA structure. The service dashboards pull from their specific RMM and PSA configuration. Nothing was adapted from a generic template or an existing product.
We build the same way for every engagement. The starting point is always your workflows, your systems, and the specific problems your team is working around every month.
Design Philosophy
This platform reflects the way we approach production software. Automation should be aggressive where it saves time and reduces error. Production actions (writing to billing systems, updating customer records, modifying quantities) require human oversight and an audit trail.
The natural-language querying capability is constrained by design. It allows staff to ask operational questions in plain language, but it cannot trigger writes or make changes. The system is clear about what it can answer versus what requires a controlled workflow.
This is not a conservative approach to automation. It is a realistic one. Software that automates correctly and maintains human control at the right points gets adopted and trusted. Software that automates without oversight gets disabled after the first billing error.
Your Workflow
Your platform would look different. That's the point.
This platform was built around one organization's specific systems, business rules, and operational workflows. We'd build yours the same way — starting from your actual processes, not from a template.