Touchless AP Automation at a US Telecom MVNO — 36K Invoices, $1.5M Savings, 95% Error Reduction
DocuMage + Workflow Automator + Workflow Planner handling 36K+ invoices across a 90K-retail-location operating footprint — with header-and-line-level extraction, archive-as-per-local-legal-requirements, and 10-week deployment.
The challenge
The client — a US-based mobile virtual network operator with approximately 90,000 retail locations providing services across the whole of the United States, operating across agreements with the largest US wireless network carriers — was running an invoice-processing workflow that the head of finance described as structurally unsustainable. The MVNO model meant that invoices arrived from the underlying network-carriers in significant volume (the wholesale-network-usage invoices, the per-subscriber service-component invoices, the per-retail-location settlement invoices, the regulatory-fee invoices) alongside the standard supplier-invoice base across the MVNO's own operational scope.
The mechanical workflow was completely manual with repetitive steps, involved the use of multiple applications, required coordination across multiple departments (the AP team for the invoice-handling, the carrier-relationship team for the wholesale-invoice validation, the regional finance teams for the retail-location settlement, the regulatory-affairs team for the regulatory-fee handling), and required high-degree-of-accuracy across 20-30 fields of data per invoice. Prior to automation, the AP department read and entered invoices, sent them for approval, waited for approval, then in case any discrepancy arose, waited for the discrepancy resolution, manually placed the invoices in physical hold-drawers until they were finally ready to be validated and paid.
The CFO's brief included specific objectives: the manual tracking-management-and-data-entry approach was slow, confusing and characterised by high error rates; the workflow needed to be transformed to support the MVNO's growth ambitions; and the regulatory-archiving requirements per local-jurisdiction had to be satisfied without imposing additional manual workflow burden.
The approach
MindMap deployed an AP-automation platform composed of DocuMage (Dm) for the multi-format invoice extraction, Workflow Automator (Wa) for the workflow-and-system-integration, Workflow Planner (Wp) for the approval-routing-and-discrepancy-resolution and Compliance Engine (Ce) for the per-jurisdiction-archive-requirements enforcement. The solution included solution-design by CPA-and-CA-and-process-experts and a consumption-based pricing model.
Phase one was the AP-process re-engineering work. We worked with the AP-team's accounts-leadership to not only identify and map out the existing workflows but also re-engineer the existing process with SMEs across the finance domain. The re-engineering identified the quick-wins for automation (the standard-invoice-format processing, the standard-approval-routing) alongside the more complex automation requirements (the carrier-wholesale-invoice validation, the retail-location-settlement reconciliation, the regulatory-fee-handling).
Phase two was the touchless-invoice-processing workflow build. DocuMage handles the invoice data-extraction across the header-level fields (invoice number, vendor, date, total amount, payment terms) and the line-level fields (per-line product description, quantity, unit price, line total). The extraction produces the structured invoice-data with per-field confidence scoring.
Phase three was the approval-and-discrepancy-resolution workflow. Workflow Planner orchestrates the approval-routing per invoice-type (the carrier-wholesale invoices route to the carrier-relationship team, the retail-location-settlement invoices route to the regional-finance teams, the standard-supplier invoices route to the standard-AP-approval path). Notification-and-alerts handle the approval-action-points; discrepancy-resolution raises escalations to the appropriate-specialist with the supporting context.
Phase four was the archive-and-compliance layer. Compliance Engine enforces the per-jurisdiction-archive-requirements (the document-retention-period requirements, the document-format requirements, the metadata-tagging requirements) with the automated archive-workflow that satisfies the requirements without imposing manual workflow burden.
The pre-built building blocks
Rather than commission a ground-up build, the engagement leaned on MindMap's pre-built accelerator library — production-tested components that compress what would otherwise be a six-to-nine-month build into weeks.
DocuMage
Header-and-line-level invoice extraction across multi-format inputs
Workflow Automator
24/7 unattended automation for ingestion-and-data-entry
Workflow Planner
Approval-routing and discrepancy-resolution case management
Compliance Engine
Per-jurisdiction archive-and-retention enforcement
The architecture
The platform runs on the MVNO's managed cloud environment with appropriate financial-data-handling controls. The Automation Anywhere RPA platform handles the unattended-automation work; Pega's workflow engine handles the approval-routing-and-case-management work.
DocuMage's invoice-extraction pipeline supports the multi-format invoice inputs (PDF, Excel, fax-converted-to-PDF). The header-level-and-line-level extraction produces the structured invoice-data with per-field confidence scoring; the extraction model has been fine-tuned on the MVNO's specific carrier-and-supplier invoice corpus for the carrier-specific format-handling.
Workflow Automator's automation workers handle the invoice-and-related-systems integration. The workers run on a 24/7 around-the-clock cadence handling the inbound-invoice ingestion, the data-validation-against-source-systems, the data-entry-into-the-AP-system and the archive-handling per the compliance requirements.
Workflow Planner's approval-routing-and-discrepancy-resolution uses Pega's workflow engine for the case-management layer. The per-invoice case maintains the lifecycle-state across the approval-routing, the discrepancy-resolution and the payment-release. The notification-and-alert workflow handles the per-approver action-points with the appropriate-escalation routing for stale-or-unresolved cases.
Compliance Engine's per-jurisdiction-archive-requirements enforcement maintains the per-state-and-per-locality regulatory rule library. The archive-workflow tags each invoice with the appropriate-jurisdiction-archive-metadata and routes to the correct archive-zone with the appropriate-retention-period configuration.
The audit trail captures every invoice-lifecycle event from ingestion through approval, payment and archive with the full context preserved for the financial-and-regulatory-audit requirements.
The numbers behind the story
Approximately 36,000 invoices have been processed through the platform across the operational period, with $1.5M projected savings over the 3-year horizon. The savings combine the direct labour-cost reduction, the early-payment-discount-capture improvements and the reduced-error-rework cost.
Errors have reduced 95% through the structured extraction-validation-and-posting workflow. The per-invoice defect rate has dropped to negligible levels; the residual defects are dominated by genuinely-ambiguous cases that route to the human-review workflow.
Time-saved against the old process has reached 70% per invoice. The 24/7 around-the-clock processing means that invoices arrive and are processed without the working-hours-constraint of the previous manual workflow; the cycle time from invoice-receipt to payment-readiness has compressed substantially.
Average cost per invoice post-RPA-implementation has reduced to approximately $4.20, against the previous baseline that the CFO had identified as a structural-cost concern. The data-extraction accuracy reached 91% within 10 weeks of deployment, supporting the rapid scale-to-production rollout.
HR forms cataloguing has improved the knowledge base — the platform's structured data-handling has supported the broader employee-and-supplier knowledge-base improvements that the HR-and-finance-shared-services team had been pursuing.
Return on investment was achieved from Day 1 of automation going live — the structured savings exceed the platform-operational cost from the first operational period.
“Our manual invoice-processing was slow, confusing and characterised by high error rates — and our ninety-thousand retail-location footprint meant the workflow was structurally unsustainable. MindMap delivered a touchless AP workflow in ten weeks with one-point-five-million-dollar projected savings, ninety-five per cent error reduction and seventy per cent time-saved per invoice. Return on investment was achieved from day one.”— Chief Financial Officer· US Telecom MVNO
Why MindMap was chosen
The MVNO had previously evaluated three specialist AP-automation vendors. Two proposed platform-replacement programmes that would have disrupted the existing AP-system estate; one proposed an outsourcing-led model that the CFO concluded was not compatible with the in-house finance-team capability the MVNO had been building.
MindMap's accelerator-composition approach — bringing DocuMage, Workflow Automator, Workflow Planner and Compliance Engine around the existing AP-system estate — was the structural differentiator. The 10-week deployment timeline and the consumption-based pricing model aligned with the MVNO's preference for rapid time-to-value and outcome-linked commercials.
Our embedded telecom-finance expertise on the delivery team (two former US-telecom-CFO-organisation finance leads and a former MVNO-operations specialist) was the third factor. The CFO valued the team's understanding of the MVNO-specific invoice-types (the carrier-wholesale invoices, the retail-location-settlement invoices, the regulatory-fee invoices) rather than the team approaching the engagement as a generic AP-automation problem.
Related deployments
Telecom WhatsApp Self-Service
ChatNext deployed across SIM, billing, and bundle management — bilingual and integrated with the carrier's BSS stack — deflecting 44% of inbound contact.
AI Voice Agent for Collections
An Arabic-English voice agent replaced the outbound collections dialler, reducing cost per contact by 58% while increasing same-call promise-to-pay rate.
Voice-First African Self-Service
Voice Bot + ChatNext reached 22M subscribers in 6 local African languages — 48% of voice-based customer interaction now AI-handled.
Want an outcome like this?
Start with a 2-week AI Readiness Sprint. We deliver a prioritised use-case backlog and business case grounded in what's actually buildable with our accelerator library.