Case study / Mid-market · Accounting

Turning Check Images and Bank Statements Into Usable Data

The check reader alone turns a 2 to 10 minute search into a few seconds. Multiply that across hundreds of checks and the math speaks for itself. The bank statement parser removes 15 minutes of manual data entry per statement, and the firm processes hundreds of these per year. These tools are taking some of the most tedious work in the office and making it disappear.

The result~15 min

saved per bank statement

Check identification reduced from 2 to 10 minutes down to seconds
Bank statement transcription eliminated entirely (15 minutes saved per statement)
Both tools auto-detect matching PDFs when they appear in Google Drive
Accountants spend less time on data entry and more time on actual accounting work

01 / Before and after

The same work, two ways.

The manual wayWith Bell AI

Pull up the bank's check export PDF.

Drop the bank PDF into the watched folder.

Scroll through pages of scanned check images looking for a matching number and amount.

Let the parser highlight the target check and surface the vendor instantly.

Cross-reference the vendor against the QuickBooks transaction by hand.

Open the auto-generated Excel file for the matching bank statement.

Transcribe bank statement totals into a spreadsheet line by line.

Move straight into reconciliation instead of data entry.

Before2–10 min

Per check, scrolling the bank's check PDFs

AfterSeconds

Tool highlights the right check on screen

02 / The work

Where it stuck. What we built.

The challenge

Accountants at a Westchester CPA firm were spending 2 to 10 minutes per check manually matching handwritten check images to transactions in QuickBooks. The bank exports a PDF with all the check images, but finding the right one is a grind. You see a check number and amount in QuickBooks, but the vendor name is missing. So you scroll through pages of scanned checks trying to match it. With active clients writing hundreds of checks per year, this added up fast. On top of that, monthly bank statements were being transcribed by hand into spreadsheets, taking about 15 minutes per statement.

The solution

We built two programs. The first reads the bank's check export PDFs and highlights the relevant check immediately, pointing the accountant straight to the vendor information they need. The second reads the bank statements and outputs them as formatted Excel files, cutting out the manual transcription entirely. Both tools detect when new PDFs arrive that match the expected format.

03 / How we built it

Step by step.

Each step was tested against the team's real work before the next one started. A person reviews anything that matters before it goes out.

01Step 1 of 6

Analyzed the structure of the bank's check export PDFs to understand how check images and data are organized

02Step 2 of 6

Built a parser that extracts check identifiers and highlights the target check on screen

03Step 3 of 6

Analyzed the bank's statement PDFs and mapped every data field to a structured format

04Step 4 of 6

Built a second parser that converts bank statements into formatted Excel output

05Step 5 of 6

Added detection logic so both tools recognize new matching PDFs automatically

06Step 6 of 6

Tested against real client documents across multiple account types and statement periods

Tools

Python
Google Drive
Excel

In their words

“Matching checks used to be the worst part of my week. Now I just pull it up and the answer's right there.”

Staff Accountant · CPA Firm, Westchester, NY

Next in mid-market

→

Migrating and Restructuring a CPA Firm's Digital Workspace

Mid-market · Accounting

The outcome

100% of client folders standardized

A good place to start

What's getting in the way?

Tell us what's slowing your team down.
We'll help you find the first thing worth fixing.

80% of payroll time eliminated for Sylvia's Magic Cleaning. Read the case study

Book a free auditA conversation. A clear next step. No obligation.