Case file · September 28, 2026
Smallpdf's chargeback problem — and the machine that won back $55K
A file-conversion company was writing off disputed payments. An AI evidence machine turned the disputes into a system — and the result is measured in recovered dollars and returned hours.
protecteddocumented
- The money
- $55K recovered + $45K saved (6 months)
- Built by
- Chargeflow
- Deployed at
- Smallpdf
- Tools
- Chargeback automationBraintree / Stripe / Adyen / Recurly integrations
What was built
Chargeflow is a chargeback-recovery service: it collects evidence for each disputed payment, enriches the case, writes a response tailored to the payment provider, submits it through the provider's workflow, and tracks the outcome. Smallpdf — the file-conversion company — deployed it across the payment rails it already used.
How it works
- Each dispute is assembled automatically: transaction data, customer history and usage evidence are collected instead of being re-gathered by hand.
- The response is written per provider — Braintree, Stripe, Adyen and Recurly each have different evidence formats and rules — and submitted through the provider's own workflow.
- Outcomes are tracked so the system learns which arguments and evidence win, and the queue is prioritised by recoverable value.
The result
Over six months, Smallpdf recovered $55K in disputed revenue and saved an estimated $45K in operational effort — more than 1,100 hours — for about $100K of measurable value.
What we can’t verify
The figures come from Chargeflow's own customer case study; there is no independent audit, and the service fee is not public. Treat the numbers as vendor-published, not as verified receipts.