20+ Suppliers, Zero Manual Entry

Executive Summary

Trail Sports, a Canadian outdoor recreation retailer, was drowning in manual invoice processing. After a failed automation attempt with another consultant and facing quotes from bookkeeping agencies that would cost a part-time salary for work that should take days, they turned to WorkLoopie.

We built a custom AI-powered automation that intelligently parses PDF invoices from 20+ suppliers, handles complex Canadian tax calculations, matches vendors to QuickBooks records, and creates accurate bills, all without manual intervention.

The Challenge

Patrick Gallagher, the owner at Trail Sports, was spending hours every week on tedious invoice data entry. As he put it: "That work is really tedious and I don't like it. It's not that it's impossible to do. It's just I hate doing it. It's like homework or chores."

He had already tried to solve this problem—twice—and both attempts failed.

Failed Attempt #1: Another Automation Consultant

Patrick initially hired an automation consultant to build a Make.com workflow using Parseo for PDF parsing. The solution had multiple critical failures:

• Inconsistent extraction: Parseo would look at the same invoice twice and produce different results each time

• Vendor matching failures: Small naming differences between invoice headers and QuickBooks records caused constant errors (e.g., "Amer Sports" vs "AmerSports Canada Sales Company")

• Tax calculation chaos: When Patrick opened bills in QuickBooks, the numbers would mysteriously change and recalculate incorrectly

• Date format confusion: Different suppliers used different date formats (YYYY-MM-DD, DD-MM-YYYY), and the automation couldn't handle the variation

"QuickBooks didn't know how to deal with 'pretty good'; It needed 100%."

— Patrick Gallagher, Trail Sports

Failed Attempt #2: Bookkeeping Agency

Patrick also explored hiring an outside bookkeeping agency to handle the data entry manually. The quote came back at essentially a part-time salary, for work that should realistically take a day or two to complete each month. The economics simply didn't make sense.

The Technical Complexity

The invoice processing challenge wasn't simple. Trail Sports' 20+ suppliers each had their own invoice formats, and the Canadian tax rules added layers of complexity:

Key Challenges:

• Variable date formats across suppliers (YYYY-MM-DD, DD/MM/YYYY, etc.)

• Freight and product categories taxed differently at 5% GST, but some products (like energy bars classified as groceries) at 0%

• Vendor names on invoices rarely matched QuickBooks records exactly

• Inconsistent invoice layouts across suppliers (though each supplier was internally consistent)

• Tax accuracy was critical. GST paid on invoices directly offsets taxes owed to the government

As Patrick explained: "The taxes need to be entered correctly because the GST that we pay gets one-to-one deducted off money that we would otherwise pay the government. Reading that field correctly and getting that into QuickBooks correctly is of paramount concern."

The Solution

WorkLoopie designed and built a comprehensive invoice processing automation that solved every challenge Trail Sports faced, including problems the previous consultant couldn't crack.

Intelligent Vendor Matching

Rather than requiring exact name matches, the solution uses AI to intelligently match invoice vendors against the existing QuickBooks vendor list. When an invoice arrives from "Amer Sports Canada Sales Company" the AI recognizes it should map to "Amer Sports" in QuickBooks, just like a human would.

AI-Powered Data Extraction

An AI agent analyzes each PDF invoice and extracts all critical fields:

• Vendor identification and QuickBooks mapping

• Invoice number and invoice date (handling any date format)

• Payment terms (Net 30, Net 45, etc.) and due date calculation

• Product subtotals and freight charges (separated by category)

• GST/tax amounts with category-aware calculations

• Grand total with validation (subtotal + tax must equal total)

Built-In Validation

Before any bill is created in QuickBooks, the automation validates that the extracted numbers make sense, subtotal plus tax must equal the grand total. If something doesn't add up, the invoice is flagged for manual review rather than creating incorrect entries.

Correct QuickBooks Integration

The solution creates bills in QuickBooks using the "inclusive of tax" method with explicitly set tax amounts. This prevents the recalculation errors that plagued the previous automation. Bills stay accurate even when opened for review.

Complete Solution Architecture

• Gmail Integration → Monitors the accounting inbox for incoming invoices with PDF attachments, filtering out statements and spam

• AI-Powered Parsing → Uses OpenAI via Make's built-in AI provider to intelligently extract invoice data regardless of format

• Vendor Matching → AI matches invoice vendor names to existing QuickBooks vendors, handling naming variations automatically

• Tax Logic → Correctly handles product categories with different tax rates (5% GST vs 0% for grocery items)

• QuickBooks Online Integration → Creates bills with properly mapped fields, correct tax handling, and stable calculations

• Google Sheets Logging → Maintains audit trail with direct links to each created bill

• Gmail Label System → Automatically labels processed invoices (green) and flags items needing manual review (red)

• Email Notifications → Alerts Patrick when invoices are processed or require attention

  
  

Key Design Decisions

No Enterprise Upgrade Required: Despite Make.com support suggesting an enterprise account was needed for this complexity, WorkLoopie achieved the solution within Trail Sports' existing Make Pro subscription by using AI agents for intelligent parsing.

Built-in AI Credits: By using Make's native AI provider rather than a separate OpenAI subscription, Trail Sports avoids additional monthly costs, AI usage comes out of their existing Make operations.

Filtering for Efficiency: The automation only processes emails with PDF attachments and uses AI to distinguish invoices from statements, preventing wasted operations and keeping costs predictable.

The Results

When Patrick reviewed the completed automation, the results exceeded his expectations, particularly the ability to correctly separate freight, tax, and product subtotals that had stumped the previous consultant:

"When I looked at the results, I thought this was doing actually more than I expected it to in terms of being able to separate the freight, the tax, and the product subtotal correctly. I was quite impressed."

— Patrick Gallagher, Trail Sports

Beyond the technical success, Patrick highlighted the stark contrast with his previous experience:

"This is honestly really stress free. If this works like it's supposed to, there;s no chance that I won't find something else that I want automated."

— Patrick Gallagher, Trail Sports

Project Timeline

When we asked what this automation would give him, Patrick simply smiled and said more time outside. That’s not a small thing. At 4 hours a week, across 52 weeks, this automation gives Patrick back approximately 208 hours a year; time he can reinvest in growing Trail Sports, or spend doing what he actually loves: hiking, skiing, and getting out on his mountain bike.

Final Thoughts

Patrick’s story is one we see again and again; capable, driven business owners spending irreplaceable hours on work that adds no real value. Not because the work is unimportant, but because it should simply happen automatically.

What WorkLoopie built for Trail Sports isn’t just a technical solution — it’s 208 hours a year returned to someone who can put them to better use. Cleaner books. Accurate taxes. A back office that runs itself, while Patrick focuses on his customers, his team, and the outdoor life that drives everything he does.

The same opportunity exists for any SMB spending hours each week on repetitive, high-stakes back-office work — whether it’s invoice processing, order reconciliation, or financial reporting. The patterns are consistent across industries: ingest the data, normalize it with AI, validate it against the rules, post it to your accounting system, and alert a human only when something needs attention.

If your business is spending time on work that should be automatic, we’d love to show you what’s possible. Reach out at workloopie.com.