Kastle's $24M Series A: AI agents for lending workflows — what founders need to know about automation in collections

This article is for informational purposes only and does not constitute financial, tax, or legal advice. Consult a qualified professional for guidance specific to your situation.

Editorial note: Reviewed for accuracy by the Startup Finance Guide editorial team. Our editors cross-reference all claims against platform documentation, regulatory publications, and vendor disclosures. Last reviewed: 2026-09-22.

Kastle, a US-based AI workforce platform for consumer lending, closed a $24 million Series A round to expand its AI agent infrastructure inside financial institutions, according to Finextra, with the company reporting that its agents have already processed more than $1.8 billion in transactions across bank clients.

The raise lands at a moment when regulators and lenders are both paying close attention to how automated systems handle consumer-facing debt workflows. For founders running or building lending operations in the US, India, or Canada, the Kastle model is worth understanding not because of the company itself, but because of what it signals about where compliance risk is moving.

What changed

The traditional approach to collections automation has been robotic process automation (RPA): scripted bots that mimic human clicks inside existing software. Kastle's pitch is different. The company says its AI agents execute high-volume, repeatable workflows inside a financial institution's existing core systems, keeping systems of record current without requiring infrastructure replacement. Human staff handle complex cases that require judgment or relationship management.

This hybrid model, where AI agents and human workers operate in parallel teams, is gaining traction across the lending stack. Competitors in adjacent spaces include Floatbot, a conversational AI platform used in collections and customer service; Vodex, which runs AI voice agents for debt recovery; and Retell AI, which offers programmable voice AI for financial services workflows. Each takes a different architectural approach to the same underlying problem: how do you scale collections capacity without proportionally scaling headcount?

Kastle's specific claim, that agents work inside existing processes and controls rather than replacing them, is the detail that matters most for compliance officers. The Consumer Financial Protection Bureau (CFPB), the US regulator responsible for enforcing the Fair Debt Collection Practices Act (FDCPA) and Regulation F, has made clear that automation does not transfer liability away from the institution deploying it. Regulation F is the CFPB's Debt Collection Rule, in force since 30 November 2021, that imposes contact-frequency limits and disclosure requirements on third-party debt collectors. Those rules apply whether the contact is made by a human agent or an AI system.

The CFPB has not yet issued formal guidance specifically addressing AI agent liability in collections workflows, but its 2022 and 2023 supervisory highlights flagged automated systems as an area of active examination focus, particularly around call frequency, consent verification, and dispute handling.

What this means for founders

If you are building or scaling a lending operation, the Kastle raise is a prompt to audit your own automation stack against current regulatory requirements, not a reason to switch vendors.

Start with contact-frequency compliance. Regulation F caps telephone contact attempts at seven within seven days per debt. If your AI agent or dialer system is not tracking this at the account level and across all channels, you have a gap. This is true whether you are using Kastle, Floatbot, Vodex, Retell AI, or a homegrown system.

Next, review your vendor contracts for indemnification language. When an AI agent makes a disclosure error or contacts a consumer outside permitted hours, the question of who bears liability depends heavily on how your service agreement is written. Platforms that operate inside your existing systems, as Kastle describes its model, may argue that the institution retains operational control and therefore regulatory responsibility. Get that in writing and have counsel review it.

For founders operating across borders, the compliance picture gets more complicated. In India, the Reserve Bank of India (RBI) issued its Digital Lending Guidelines in 2022, which require that all borrower-facing communication, including automated contact, disclose the regulated entity's name and not just the technology vendor's. Canadian lenders must contend with provincial debt collection statutes that vary by territory, and the Financial Consumer Agency of Canada (FCAC) has signaled interest in how AI systems handle dispute resolution.

The operational checklist for any founder evaluating AI agent platforms in collections:

• Confirm the platform logs every consumer contact attempt with timestamp, channel, and outcome, in a format your compliance team can audit.
• Verify that dispute and cease-communication requests trigger immediate suppression across all channels, not just the one where the request was received.
• Ask the vendor how their system handles state-level or provincial variations in contact rules. A federal-only compliance posture is not sufficient in the US or Canada.
• Understand the data residency model. If your borrower data is processed by an AI agent running on third-party infrastructure, that may trigger obligations under the Foreign Exchange Management Act (FEMA) for India-linked entities or under Canada's Personal Information Protection and Electronic Documents Act (PIPEDA).

Limitations and open questions

Kastle's $1.8 billion in processed transactions is a notable figure, but the company has not published independent audit results, third-party compliance certifications, or details about which institutions are using the platform. The Finextra report does not name any bank clients, and Kastle's public documentation does not specify which core banking systems its agents integrate with. Founders evaluating the platform or its competitors should request reference clients in their specific regulatory jurisdiction before signing.

More broadly, the regulatory treatment of AI agents in collections is still forming. The CFPB has not published formal guidance on whether an AI agent constitutes a "debt collector" under the FDCPA when it operates under the supervision of a creditor rather than a third-party servicer. That distinction matters for which rules apply. A 2024 Reuters report on CFPB enforcement priorities noted that the bureau was examining automated systems in servicing contexts, but no enforcement action specifically targeting AI agent vendors has been made public as of this writing.

The Federal Trade Commission (FTC) has separate authority over unfair or deceptive acts and practices, and its 2023 policy statement on AI made clear that existing consumer protection law applies to automated systems. But neither the FTC nor the CFPB has issued a rule that specifically addresses the hybrid human-AI team model that Kastle and its competitors are selling.

For founders, that uncertainty is the point. The technology is moving faster than the regulatory guidance. Building your compliance posture around current rules, while monitoring CFPB and FTC rulemaking, is the only defensible approach right now.

This article is for informational purposes only and does not constitute financial, tax, or legal advice. Consult a qualified professional for guidance specific to your situation.