With Chargeback Agent, we’ve transformed how we manage disputes. Now, we’re recovering more dollars with less labor, and we can efficiently handle chargebacks for our agencies even as we continue to grow rapidly.
—Mark Klinker, AllPaid CFO
Managing chargeback workflows on behalf of government agencies
AllPaid is a leader in government technology, processing tax, utility, and corrections payments for more than 3,500 municipalities and counties.
As part of its mission to reduce the cost and friction of payment acceptance, AllPaid, which is a payment facilitator, takes responsibility for chargebacks on its platform on behalf of its agencies. This valuable customer benefit allows the agencies to focus on delivering services to their constituents rather than completing burdensome chargeback workflows.
At the same time, this commitment entails significant labor costs and chargeback losses for AllPaid. In fact, under its previous ownership, AllPaid did not fight chargebacks less than $200 due to the labor costs involved, meaning these chargebacks—even for legitimate government transactions—were conceded and impacted AllPaid’s bottom line.
Increasing efficiency and improving win rates with an AI‑native approach
In 2025, AllPaid was acquired by private equity investors Theatre Capital and LoneTree Capital. Following this transaction, AllPaid’s new owners brought in forward‑thinking executives to lead the next chapter of the company’s growth. These leaders knew they wanted to multiply the revenue scale of AllPaid without multiplying the OpEx associated with the chargebacks function.
AllPaid partnered with Findustry AI to launch Chargeback Agent in Fall 2025. Chargeback Agent works like a skilled human analyst, collecting evidence from merchant systems, interpreting card network requirements, building persuasive arguments in a rebuttal letter, and submitting the package via processor APIs for the issuing bank’s review. And, because the AI agent is directly integrated to processor APIs, it does not need to be prompted by the user to start its work; rather, Chargeback Agent is an “ambient agent,” meaning it automatically prepares responses as soon as a chargeback occurs.
In addition to faster resolution of chargebacks, Chargeback Agent also optimizes win rates by incorporating evolving best practices in the rapidly changing chargebacks space. For example, Chargeback Agent now delivers further savings to AllPaid by aligning evidence to Compelling Evidence 3.0 requirements, fighting chargebacks that are procedurally invalid (i.e., inconsistent with network rules), and geolocating the payer’s IP address in cases of first‑party fraud.
As a result, AllPaid’s recovered dollars as a percentage of chargeback volumes have increased from an average of 11.83% prior to deploying Chargeback Agent to as much as 36.1%—a relative increase of 205%—with Chargeback Agent, representing six figures of savings per year.
Chargeback Agent Win Rate vs. Prior Baseline
Win Rate (%)
Recovered dollars as a percentage of chargeback volumes, by month after deploying Chargeback Agent, compared against AllPaid’s prior baseline of 11.83%. Win rate increased by +205% from that baseline by Month 3.
And, with less time spent handling chargebacks, AllPaid team members are now able to take on higher‑leverage tasks for the business, such as customer communications and compliance functions related to AllPaid’s state regulatory requirements. Had the existing team members not been freed up for these tasks, AllPaid would have had to hire a compliance manager and add to its labor cost.
Based on these outcomes, Findustry AI founder and CEO Jonathan Razi commented, “AllPaid’s successful example shows that other growth companies can strategically use GenAI to keep headcount steady and accomplish even more with their existing team.”
AllPaid’s successful example shows that other growth companies can strategically use GenAI to keep headcount steady and accomplish even more with their existing team.
—Jonathan Razi, Findustry AI Founder & CEO
Integrating with existing infrastructure while expanding capabilities
AllPaid selected Chargeback Agent in part because many legacy chargeback products offer rigid templates built for retail eCommerce, which are poorly adapted to other verticals such as consumer‑to‑government payments. Instead, AllPaid wanted a system that could incorporate their merchant and industry context into dynamic responses, making purpose‑built generative AI a strong fit.
Findustry AI’s forward‑deployed engineers were able to set up all the required connections to existing AllPaid internal systems and payment processor endpoints with minimal lift on the AllPaid side: in fact, AllPaid’s technical team wrote zero new code for the implementation, allowing R&D resources to remain focused on other innovation priorities and achieve a faster time to value.
Finally, AllPaid wanted an AI agent that could be supervised and managed by their team, rather than fully outsourced with little transparency. In the Findustry AI portal, AllPaid is able to configure which chargebacks are automated versus drafted for human review based on thresholds such as dollar amount.
For those select cases, Findustry AI’s human‑in‑the‑loop interface enables the AllPaid team member to review and approve the AI agent response or upload additional evidence, reducing the time spent handling nuanced cases from hours to minutes or even seconds. A built‑in audit trail captures all AI agent actions as well as human edits or approvals, and this usage data powers continuous improvement of the AI’s responses.
Partnering for long‑term AI transformation including business intelligence
Chargebacks should not exist in a silo. Best‑in‑class teams are able to analyze their chargebacks for valuable insights into their overall payments strategy and learn from trends related to fraud, customer support, product or service quality, and processing practices (such as authorizations and recurring payments) that may affect their business. In the past, however, outsourcing chargebacks to legacy vendors has often meant reduced visibility into these trends.
Next, AllPaid will be leveraging Findustry AI’s Insights Agent, which connects to their data via MCP (Model Context Protocol), to better understand their chargebacks and payments trends. Instead of clicking through processor portals or exporting raw data to Excel, the AllPaid team can now query their own data in natural language and quickly create custom charts and reports without needing additional resources.
Ready to start your AI transformation?
If we can help your organization, reach out to Findustry AI today.