OpenDebt, a Sydney-based startup building artificial-intelligence systems for lenders and debt-collection agencies, has raised A$2 million in pre-seed funding from Blackbird Ventures and Antler, betting that carefully governed AI agents can transform one of finance's most sensitive and labour-intensive functions.
The round was announced on 22 September 2026. OpenDebt was founded in 2025 by Norman Yan, a former banker at Westpac, one of Australia's largest banks, and Jonas Tischer.
Agents that talk, and agents that assist
OpenDebt's technology can operate in two modes. It can assist human collectors, helping them prioritise accounts, prepare conversations and handle administrative tasks. Or it can communicate directly with borrowers through voice calls, SMS and email.
Before discussing any account, the system verifies the identity of the person it is speaking with, a critical safeguard in an industry where disclosing debt information to the wrong person can breach privacy rules and cause harm. Cases involving financial hardship, sensitive circumstances or a request to speak with a person are routed to human specialists.
That design reflects a recognition that debt collection is not merely a matter of efficiency. It involves people who may be under financial stress, dealing with illness, job loss or family difficulties. Handled poorly, collections can cause significant distress and expose lenders to regulatory and reputational risk.
Why the economics favour automation
The opportunity OpenDebt is targeting is particularly large in smaller delinquent balances. For many lenders, the cost of assigning human collectors to pursue small overdue amounts can exceed the amount likely to be recovered. As a result, such balances are often left unpursued, written off or sold at steep discounts to debt buyers.
AI agents could change that calculation. Automated systems can contact large numbers of borrowers at low marginal cost, offer flexible repayment options and follow up consistently, making it economically viable to engage with accounts that would otherwise be neglected. For borrowers, a more accessible, less confrontational channel could make it easier to arrange manageable repayment plans before debts escalate.
Governance as product
The challenge is that financial AI is increasingly a governance problem as much as a technology problem. An autonomous agent must know not only what to say, but when to stop acting. It must respect customer-specific policies, consumer-protection rules, accurate balances, documentation requirements, identity verification and escalation rules.
In Australia, debt collectors must comply with guidance jointly issued by the Australian Securities and Investments Commission and the Australian Competition and Consumer Commission, which sets out expectations on matters such as contact frequency, harassment and dealing with vulnerable consumers. Lenders also have legal obligations to consider hardship requests. In the United States, where OpenDebt plans early partnerships, debt collectors are governed by the Fair Debt Collection Practices Act and the Consumer Financial Protection Bureau's Regulation F, which include detailed rules on communications.
The value of OpenDebt's product therefore depends heavily on encoding these constraints reliably around its AI models. A system that is persuasive but careless would be a liability; one that is compliant, respectful and effective could become a valuable tool for lenders.
Early traction
OpenDebt is already expanding deployments with Australian lenders and collection agencies, according to the company, and plans to establish early partnerships in the United States. The US market is considerably larger, with a vast consumer-credit sector and a well-developed collections industry, but it also brings a more complex, litigation-prone regulatory environment.
Expanding internationally at the pre-seed stage is ambitious. It will require the company to adapt its compliance logic to different legal frameworks, build relationships with a new set of lenders and agencies, and demonstrate results that justify adoption.
Blackbird and Antler's bet
Blackbird Ventures is one of Australia's most prominent venture firms, having backed many of the country's best-known technology companies. Antler, a global early-stage investor that supports founders from the idea stage, has built a portfolio across multiple continents. Their joint backing gives OpenDebt both local credibility and a network that could support international expansion.
For both investors, OpenDebt fits a wider thesis: vertical AI applications that target specific, high-value workflows in regulated industries. Rather than building general-purpose AI, such companies apply models to narrow domains where they can accumulate specialised data, embed regulatory knowledge and deliver measurable results.
The ethical dimension
The use of AI in debt collection raises understandable concerns. Critics worry that automated systems could contact borrowers too frequently, apply pressure without empathy or fail to recognise vulnerability. There are also questions about transparency: borrowers may want to know whether they are speaking with a human or a machine.
OpenDebt's emphasis on identity verification and human escalation addresses some of these concerns, but the company will need to demonstrate through its practices that automation improves, rather than degrades, the experience of people in financial difficulty. Clear disclosure, respectful communication and easy access to human support will be essential.
The founders' banking perspective
Norman Yan's background at Westpac gives OpenDebt an insider's understanding of how large lenders manage collections, compliance and customer outcomes. Banks operate under intense regulatory scrutiny in Australia, particularly following the Royal Commission into misconduct in the financial services industry, which highlighted poor treatment of customers across the sector. That experience is likely to shape how OpenDebt designs its product for institutional clients, who will demand detailed audit trails, configurable policies and evidence that automated interactions meet conduct standards.
For lenders, the appeal of such a system lies in consistency. Human collectors vary in their skill, tone and adherence to scripts, and supervising large teams is costly. A well-governed AI system can apply the same policies to every interaction, log every conversation and flag issues for review, potentially improving both compliance and customer outcomes.
What comes next
With A$2 million in fresh capital, OpenDebt's immediate priorities are likely to include expanding its Australian customer base, strengthening its compliance and escalation systems, and laying the groundwork for its US entry.
The company is part of a broader wave of startups bringing AI agents into financial operations, from payments and lending to accounting and audit. In each case, the defining question is the same: can software be trusted to act, not just advise, in environments where mistakes carry real consequences? OpenDebt's answer is that trust depends on design, and that the best AI systems in finance will be those that know their own limits.
