From Paper Invoices to AI-Powered Insights: How Loop8 Transformed MGA Liquor's Multi-Store Operations with AWS
MGA Liquor, a family-owned chain of seven retail liquor stores in California's Central Valley, partnered with Loop8 LLC to replace paper-based invoicing and disconnected legacy systems with a cloud-native platform on AWS — reducing invoice processing time by 85%, recovering 20 hours of staff time per week, and identifying an estimated $280,000 in annual savings from shrinkage and dead stock reduction.
Faster invoicing
From 15 min to under 2 min per invoice
Staff time recovered
Redirected from data entry to customer service
Annual savings
Shrinkage detection and dead stock reduction
First-year ROI
Return on investment achieved in year one
Executive Summary
MGA Liquor, a family-owned chain of seven retail liquor stores in California's Central Valley, was losing an estimated $350,000 annually to shrinkage, spoilage, and dead stock — and had no way to see it happening in real time. Paper invoices consumed more than 20 hours of staff time every week, stores operated on disconnected Windows 7 point-of-sale systems with no centralized reporting, and the business owner relied on weekly in-person visits to each location just to understand how operations were performing.
Loop8 LLC, an AWS Select Tier Partner, designed and delivered a cloud-native platform that digitized MGA Liquor's core operations from the ground up. The solution leveraged Amazon Textract for AI-powered invoice scanning, AWS Lambda and Amazon DynamoDB for serverless data processing, Amazon Bedrock for natural-language business insights, and Amazon EC2 with AWS Managed Microsoft AD for migrating legacy Windows infrastructure. Within weeks of launch, MGA Liquor saw an 85% reduction in invoice processing time, recovered 20 hours per week of staff labor, and gained real-time visibility across all seven stores — with an estimated $280,000 in annual savings from early shrinkage detection and dead stock identification alone.
About the Customer
MGA Liquor is a family-owned chain of seven retail liquor stores serving communities across California's Central Valley. The business has been in operation for more than 15 years, growing from a single storefront to seven locations, each carrying between 2,000 and 3,000 SKUs spanning spirits, wine, beer, and tobacco products. In addition to walk-in retail customers, MGA Liquor supplies local restaurants and bars through wholesale distribution, generating approximately $12 million in annual revenue and employing around 50 people.
Background
Before engaging Loop8, MGA Liquor's technology environment reflected the realities of a growing small business that had never undergone a deliberate IT modernization. Each of the seven stores operated independently with Windows 7-based POS terminals running a legacy point-of-sale application backed by SQL Server Express. There was no centralized data repository connecting the locations. Store managers exported CSV files manually from each terminal and emailed them to the owner, who reviewed them in spreadsheets — a process that introduced delays of one to two weeks before any actionable insight was available.
The back-office infrastructure was equally fragmented. An on-premises Active Directory domain controller and shared file server at the main office were approaching end-of-life. Supplier invoices arrived exclusively on paper, and each store processed an average of 25 to 30 invoices per week by manually keying line items into QuickBooks Online. The owner had no mechanism to compare performance across locations, detect purchasing anomalies, or identify products that were sitting on shelves unsold.
The Challenge
The operational cost of this fragmentation was staggering. Manual invoice data entry consumed approximately three hours per store per week — more than 20 hours total across all seven locations. Errors in manual transcription led to payment discrepancies with suppliers, missed early-payment discounts, and chronic difficulty reconciling inventory received against inventory sold. Staff members who could have been serving customers or managing store operations were instead hunched over keyboards entering line items from paper invoices.
More critically, MGA Liquor had almost no ability to detect shrinkage — whether from theft, breakage, or simple administrative error. Without automated inventory reconciliation, the gap between what was purchased and what was sold went largely unmonitored. Slow-moving products tied up working capital on shelves for months before anyone noticed. The owner estimated that across all seven stores, the business was losing approximately $350,000 per year to undetected shrinkage, expired products, and dead stock that should have been returned or discounted.
The underlying Windows 7 infrastructure compounded these risks. The POS terminals were running on an operating system well past end-of-support, creating security vulnerabilities that a small business without dedicated IT staff had no practical way to address. The on-premises domain controller and file server required hardware replacement, and the legacy IIS-hosted internal web application used for inter-store communication needed migration to a supported platform before it became a liability.
“I was driving to each store every week just to understand what was happening in my own business. By the time I saw the numbers in a spreadsheet, the data was already two weeks old. I knew we were losing money to shrinkage and dead stock, but I couldn't tell you exactly where or how much — it was all guesswork.”— Gurpreet Singh, Co-founder and President, MGA Liquor
Why Loop8
MGA Liquor selected Loop8 LLC as their implementation partner after evaluating several options. Loop8's specialization in delivering cloud solutions for small and medium businesses set the firm apart from larger consulting practices that typically design for enterprise-scale budgets. Rather than proposing a multi-vendor engagement that would have required MGA Liquor to coordinate between infrastructure, application, and AI specialists, Loop8 offered full-stack delivery — infrastructure-as-code, serverless backend, React frontend, AI integration, and Windows migration — as a single integrated engagement.
Loop8's pre-built modular Terraform architecture and serverless design patterns meant the team could move from concept to working system in just four weeks, delivering tangible value before the engagement became a prolonged consulting exercise. The promise of rapid time-to-value was a decisive factor for a family-owned business that needed results, not roadmaps.
Why AWS
AWS was the natural platform choice for this engagement. The serverless-first approach — anchored by AWS Lambda, Amazon API Gateway, Amazon DynamoDB, and Amazon S3 — eliminated the need for MGA Liquor to manage servers, a critical advantage for a small business without dedicated IT staff. AI and machine learning services like Amazon Textract and Amazon Bedrock delivered production-ready invoice OCR and natural-language analytics without requiring data science expertise or model training. The pay-per-use pricing model aligned with MGA Liquor's variable workload patterns, keeping monthly costs predictable. And for the legacy Windows infrastructure, Amazon EC2, AWS Managed Microsoft AD, and Amazon FSx for Windows File Server provided a clean migration path without requiring application rewrites — bridging the gap between modern cloud architecture and the Windows-dependent POS systems the stores relied on daily.
The Solution
Loop8 began with a rapid assessment of MGA Liquor's operations, cataloging every data flow from POS terminals to supplier invoices to QuickBooks entries. The engagement was structured in phases: first, a serverless data foundation; second, AI-powered invoice digitization; third, business intelligence and insights; and fourth, legacy Windows infrastructure migration — with enterprise-grade security woven throughout.
Centralized Data Platform
The foundation of the platform is a centralized data lake built on Amazon S3 with store-level isolation using prefix-based partitioning. Each store's POS system now exports daily sales data via a PowerShell script and the AWS CLI to the appropriate S3 prefix. AWS Glue crawlers automatically discover and catalog new data, maintaining table schemas in the Glue Data Catalog, while pre-built Amazon Athena views provide real-time analytics across sales, product performance, inventory balance, shrinkage analysis, and slow-mover identification.
AI-Powered Invoice Digitization
The invoice digitization pipeline transformed MGA Liquor's most time-consuming manual process. When a store employee photographs a supplier invoice using the progressive web app on their smartphone, the image uploads directly to S3 via a presigned URL. The upload triggers an AWS Lambda function that calls Amazon Textract's AnalyzeExpense API to extract vendor names, invoice IDs, dates, line items, quantities, unit prices, and totals — all without human intervention. Structured invoice data is stored in Amazon DynamoDB with a composite key design enabling fast retrieval by store and invoice, and a DynamoDB Stream triggers AWS Step Functions for downstream orchestration. The Step Functions workflow routes processed invoices to a QuickBooks sync Lambda that automatically creates corresponding bills in QuickBooks Online via the Intuit API, complete with retry logic and error handling.
“The first time I watched an employee snap a photo of an invoice and saw the line items appear in QuickBooks two minutes later — that's when I knew this was going to change everything. We went from a full day of data entry every week to basically zero.”— Gurpreet Singh, Co-founder and President, MGA Liquor
AI-Powered Business Insights
Two AI-powered capabilities were integrated using Amazon Bedrock. An interactive AI assistant, powered by Amazon Nova Lite, allows the business owner to ask natural-language questions about store performance through a chat interface — questions like “What were my top five products last week across all stores?” or “Show me shrinkage alerts for store three.” The assistant translates those questions into Athena SQL queries, executes them, and returns formatted insights in seconds. A second automated capability runs daily via a scheduled Lambda function, aggregating key metrics through Athena and generating a natural-language executive summary delivered to the owner's inbox each morning via Amazon SNS.
Progressive Web App
The frontend is a React progressive web app hosted on Amazon S3 with Amazon CloudFront for HTTPS delivery and edge caching. Amazon Cognito provides authentication with custom attributes for multi-store access control, supporting owner, manager, and employee roles with differentiated dashboard visibility. The PWA runs on employees' existing smartphones with no app store deployment required, and service worker caching provides resilience during intermittent connectivity at store locations.
Legacy Windows Migration
For the legacy Windows infrastructure, Loop8 migrated the on-premises domain controller to AWS Managed Microsoft AD, the file server to Amazon FSx for Windows File Server with automated seven-day backups, and the terminal server and IIS web application to Amazon EC2 instances running Windows Server 2022. AWS Systems Manager Patch Manager now applies critical and security patches automatically every week, replacing the previous irregular manual patching process and achieving full compliance.
Enterprise-Grade Security
Security was not an afterthought — it was architected into every layer. AWS Config monitors compliance across 11 managed rules with auto-remediation via SSM Automation. AWS CloudTrail captures all API activity with log validation and data event logging. Amazon GuardDuty provides continuous threat detection, AWS Security Hub aggregates findings against both the AWS Foundational Security Best Practices and CIS Benchmarks, Amazon Inspector scans Lambda functions and EC2 instances for vulnerabilities, and AWS WAF protects the API Gateway with rate limiting, SQL injection prevention, and bot control rules. All data at rest is encrypted using customer-managed AWS KMS keys, and all compute resources are isolated in private subnets within an Amazon VPC with gateway and interface endpoints ensuring traffic to AWS services never traverses the public internet.
Infrastructure as Code
The entire infrastructure is defined in Terraform across 13 reusable modules, deployed through a GitHub Actions CI/CD pipeline with automated builds, cost estimation via Infracost, and OIDC-based role assumption — no long-lived credentials anywhere in the pipeline.
Results
The impact was immediate and measurable. Invoice processing time dropped from 15 minutes per invoice with manual entry to under two minutes with automated OCR — an 85% reduction. Staff hours devoted to invoice data entry fell from more than 20 hours per week across all stores to fewer than two hours of exception review, a 90% reduction that freed employees to focus on customer service and store operations. Shrinkage detection moved from quarterly manual audits to daily automated reconciliation, and data visibility lag collapsed from one to two weeks to near-real-time.
Financially, the platform delivered an estimated $280,000 in annual savings from early shrinkage detection and dead stock identification, reducing the previous $350,000 annual loss by approximately 80%. Automated invoice processing accounts for an additional $52,000 in annual labor savings, and the elimination of on-premises server hardware maintenance saves another $15,000 per year. Monthly AWS costs run between $1,200 and $1,800 under a serverless pay-per-use model with EC2 Reserved Instances, yielding an estimated 12:1 return on investment in the first year.
Beyond the numbers, the transformation changed how MGA Liquor operates as a business. The owner now receives AI-generated insights every morning summarizing performance across all seven stores, complete with shrinkage alerts and slow-mover recommendations — without setting foot in a single location. Centralized Active Directory and shared file storage replaced per-store local accounts and USB-driven file sharing. And the modular architecture means adding a new store requires nothing more than provisioning a new store ID in the Terraform configuration — no code changes required.
Annual Savings Breakdown
- Shrinkage & dead stock reduction$280,000
- Invoice processing labor$52,000
- Server hardware maintenance$15,000
- Total estimated annual savings$347,000
Platform Economics
- Monthly AWS costs$1,200 – $1,800
- Pricing modelServerless pay-per-use
- First-year ROI12:1
- Infrastructure13 Terraform modules, CI/CD
“This isn't just a technology upgrade — it's a completely different way of running my business. I went from not knowing what was happening across my stores to having an AI tell me exactly what I need to pay attention to every single morning. The ROI paid for itself before we even finished rolling out to all seven locations.”— Gurpreet Singh, Co-founder and President, MGA Liquor
AWS Services Used
About the Partner
Loop8 LLC is an AWS Select Tier Partner specializing in cloud-native solutions for small and medium businesses. Loop8 helps SMB customers modernize their operations by leveraging serverless architectures, AI and machine learning services, and infrastructure-as-code practices that deliver enterprise-grade capabilities at cost points appropriate for growing businesses. Loop8 holds AWS Service Delivery designations for AWS Lambda, Amazon DynamoDB, AWS Config, and Amazon EC2 for Windows Server, along with the AWS Security Competency and AWS SMB Competency.
Transform Your Multi-Location Operations
See how AWS-powered automation can eliminate manual processes and deliver real-time visibility across your business.