Change Request Automation at a European MSI Integrator — 95% AHT Improvement, <2-Minute Per-Transaction
Workflow Automator + Multi-Agent Orchestrator + REST API integration handling cross-environment change-request management across 7 applications with custom multi-layered pattern-matching.
The challenge
The client — one of the largest MSI (multi-source-integrator) integrators in the world, receiving multiple change-requests from its 1000-plus enterprise customers and managing them centrally with a team of agents spread across Europe — was running a change-request management workflow that had become a structural cost-and-cycle-time bottleneck. The process had three phases: collecting data from multiple applications, validating the data and making the changes, and responding to the change-request.
The big problems were specific. Up to 4 hours were being spent per case across the workflow; the delay in responding was leading to revenue-leakage as the customers' change-requests were not being closed within the contractual SLAs; and the cost per transaction handled was very high given the per-agent time-investment per case.
The structural complexities that had defeated the previous automation-attempts were: applications across internal-plus-external customer-environments leading to difficulty in connecting with all data-sources; extraction of data from within a complex Visio LLD (low-level design) document being a major choke-point in the entire process-flow; uncertain-and-non-fixed responses from the customer-systems' APIs creating difficulty in parsing the data or writing stable code; and as many as 24 layers of responses for one of the applications creating slowness-of-response and complexity in obtaining the relevant data.
The approach
MindMap deployed an intelligent-process-automation platform composed of Workflow Automator (Wa) for the cross-environment integration, Multi-Agent Orchestrator (Mo) for the workflow coordination, REST API and .NET-scripting capabilities for the custom-integration-and-response-handling, and Computer Vision for the complex Visio LLD extraction work.
Phase one was the complexity-mapping work. Each of the structural complexities was mapped to a specific solution-pattern. The cross-environment integration was addressed through individually-built APIs for the 7 different applications over secure-gateways. The complex Visio LLD extraction was addressed through a multi-layered pattern-matching algorithm to extract all the information. The uncertain-and-non-fixed API responses were addressed through custom .NET-code to handle the multiple-response-scenarios. The 24-layer-response-application was addressed through a code-navigator-algorithm that skims through the various response-servers identifying the right one.
Phase two was the touch-free-automation build. The intelligent process-automation handles the entire process end-to-end in an unattended fashion — business-users wake up in the morning to see a pre-processed change-request at automation triggered automatically rather than the previous workflow that required the agent's manual-orchestration across the multiple applications.
Phase three was the response-handling-and-AHT-improvement work. The automated solution processes the entire transaction in under 2 minutes per transaction against the previous 4-hour baseline — a >95% improvement in AHT. The cost-per-transaction has reduced commensurately.
Phase four was the freeing-up-productive-hours work. The business can now truly focus on the business-work by taking the agent's mind off the manual-repetitive tasks, with thousands of productive-hours freed up for the customer-relationship and the strategic-engagement work that the manually-orchestrated workflow had been crowding out.
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.
Workflow Automator
Cross-environment integration with per-application APIs over secure gateways
Multi-Agent Orchestrator
End-to-end workflow coordination
DocuMage
Complex Visio LLD extraction with multi-layered pattern-matching
Workflow Planner
Custom response-handling for uncertain API patterns
The architecture
The platform runs on the integrator's managed cloud environment with the appropriate cross-environment integration patterns. The cross-environment integration spans the integrator's internal applications and the customer-environment applications across the 1000-plus enterprise customer base.
Workflow Automator's individually-built APIs for the 7 different applications use the appropriate per-application authentication-and-data-access pattern. The APIs run over secure-gateways with the appropriate access-controls and the per-customer audit-trail capture.
The custom multi-layered pattern-matching algorithm for the Visio LLD extraction handles the complex-Visio-document parsing with the structured-information extraction that the downstream workflow requires. The algorithm uses Computer Vision for the visual-element identification combined with the layered pattern-matching for the cross-element relationship-resolution.
The custom .NET-code for the multiple-API-response-scenario handling addresses the structural-uncertainty in the customer-systems' API responses. The code handles the per-response-pattern identification, the per-response-content extraction and the per-response-error-handling that the previous code-approaches had not been able to handle stably.
The code-navigator-algorithm for the 24-layer-response-application skims through the various response-servers identifying the right one. The algorithm handles the structural-slowness of the multi-layer-response pattern with the appropriate-timeout-and-retry handling.
Multi-Agent Orchestrator coordinates the end-to-end workflow with the per-change-request lifecycle tracking. The audit trail captures every change-request-lifecycle event with the full context preserved.
The numbers behind the story
AHT improvement of >95% has been achieved across the change-request workflow. The previous 4-hour-per-case time-investment has compressed to under 2 minutes per transaction through the touch-free automation pattern.
Revenue-leakage from the delayed-response cycle has been eliminated. The change-requests now close within the contractual SLAs with substantial headroom, removing the revenue-leakage exposure that the manually-orchestrated workflow had been generating.
Cost-per-transaction has reduced substantially through the combination of the per-transaction time-reduction and the per-transaction agent-attention-reduction. The cost-per-transaction economics now support the integrator's profitable-growth on the change-request workflow rather than being a structural-cost-drag.
Productive-hours have been freed up across the agent-base. The business can now truly focus on the customer-relationship and the strategic-engagement work that the manually-orchestrated repetitive-task workload had been crowding out.
An unexpected outcome: the structured workflow has surfaced customer-behaviour patterns that were invisible in the manually-orchestrated workflow. The integrator's customer-success team is now using these patterns to drive the customer-engagement-strategy work that the previous workflow had structurally prevented.
“Our change-request workflow had been a structural cost-and-cycle-time bottleneck that two previous RPA vendors had been unable to crack. MindMap addressed each structural complexity with the appropriate solution-pattern — individual APIs, custom pattern-matching, custom response-handling, the code-navigator algorithm. Ninety-five per cent AHT improvement, transactions processed in under two minutes, and our agent-base freed up for the customer-engagement work that matters.”— Head of Service Operations· European MSI Integrator
Why MindMap was chosen
The integrator had previously attempted the automation work with two specialist RPA vendors. Both had been unable to handle the structural complexities (the cross-environment integration, the complex Visio LLD extraction, the uncertain API responses, the 24-layer response handling) that had defeated the previous attempts.
MindMap's accelerator-composition approach — bringing Workflow Automator's cross-environment integration capability together with the custom .NET-scripting, the REST API integration, the Computer Vision, the Visual Basic and the code-navigator algorithm work — was the structural differentiator. The composition addressed each structural complexity with the appropriate solution-pattern rather than attempting a one-size-fits-all approach.
Our embedded enterprise-integration expertise on the delivery team (two former enterprise-integration architects from peer MSI providers and a former .NET-architect with REST-API depth) was the third factor. The integrator's automation leadership valued the team's ability to engage with the structural-complexities as engineering problems with concrete solution-patterns rather than as abstract obstacles.
Related deployments
Background Checks at Scale
A unified RPA + IDP stack handling identity, employment, education, criminal and credit checks across 40 countries — 3,000 checks per day, straight-through.
RPA-to-Agentic Migration
A migration from 38,000 brittle RPA bots to 240 agentic workflows — handling more volume at a fraction of the maintenance burden.
Big-4 AI Delivery Practice
MindMap's platform underpins the Big-4 firm's AI delivery practice — 2,400 consultants delivering AI engagements to the firm's enterprise customer base.
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.