Today, we’re launching PathPilot Agents for Lending: specialized AI agents that help consumer lenders automate origination, servicing, collections, and quality on their operations.

The problem

Consumer lenders need to grow, but lending operations are still expensive, manual, and dependent on large internal teams and external vendors.

Borrowers are often asked to submit the same information multiple times, wait for manual reviews, and navigate disconnected processes across origination, servicing, and collections.

For lenders, this means higher operating costs, slower decisions, inconsistent customer experiences, and difficulty scaling without adding more headcount.

Current AI solutions are not enough. Without specialization, guardrails, and the right operational context, agent quality breaks down—especially in regulated, customer-facing workflows.

What we built

PathPilot builds specialized AI agents for lending operations, powered by a proprietary harness that connects AI models directly to each lender’s workflows, systems, policies, and data.

It starts with the Origination Agent, which guides borrowers through WhatsApp or web forms, collects the right information, pre-verifies documents, and connects with underwriting, identity, and fraud systems to support faster decisions.

Once the loan is active, the Servicing Agent helps customers with payments, balances, questions, and account updates across voice, WhatsApp, chat, and email.

When a payment is missed, the Collections Agent follows up compliantly, keeps full conversation and account context, connects with call center and CRM systems, and negotiates next steps to drive resolution.

Behind every workflow, the Audit and Quality Agent reviews every interaction and action, identifies quality or compliance risks, and provides actionable feedback that helps both human teams and AI agents improve continuously.

https://www.youtube.com/watch?v=aAQ1mLB-qPo

One agent or a coordinated workforce

Lenders can start with a single agent for one workflow or deploy a coordinated workforce of agents that work together and share context.

For example, the Origination Agent can pass a customer’s application history to the Servicing Agent. If the account later becomes delinquent, the Collections Agent can use the same operational context rather than starting the relationship from zero.

The Audit and Quality Agent works across the entire system, creating a continuous feedback loop that improves accuracy, customer experience, and compliance.

Because PathPilot is built on a modular framework, internal AI and engineering teams can extend, customize, and ultimately own the system as their capabilities grow.

We do not want lenders to become permanently dependent on another black-box vendor. We want to give them a reliable foundation for building and operating their own AI workforce.

Why we built this

The problem became personal when I tried to finance my first car after moving to the U.S.

I had no local credit history and very little understanding of how the process worked. I went through a long and frustrating process in which multiple lenders manually reviewed my application, repeatedly requested information, and ultimately offered me a loan with very high fees.

Today, as a real estate investor, I still experience many of the same problems when applying for a mortgage or refinancing a property. The process remains slow, repetitive, and difficult to navigate—even when you understand how lending works.

There needs to be a better way.

Before PathPilot, David and I spent years building and operating AI first products at Meta. We learned that making AI work reliably in production requires much more than choosing a model. It requires integrations, evaluations, guardrails, feedback loops, and deep operational context.

We started PathPilot to bring that infrastructure to lending—so lenders can operate more efficiently and borrowers can receive a faster, clearer, and more responsive experience.

Where we are today

PathPilot agents are already supporting fintech lenders, BNPLs, embedded lending platforms, and collections organizations across the U.S. and Latin America.

Our customers use PathPilot for workflows including conversational origination, document validation, servicing, collections, and quality assurance.

Our ask

We would love to speak with:

• Consumer lenders, BNPLs, neobanks, and embedded lending companies
• Operations, servicing, collections, risk, and compliance leaders
• Lending platforms interested in offering specialized AI agents to their customers
• Internal AI teams looking for a framework they can customize and eventually own

Book a demo at getpathpilot.com.