An AI agent startup aiming to replace manual bank underwriting

This article is a part of Startup Spotlight, a series that features young, up-and-coming startups.

Image credit: Timmy Loen

Sergei Kuzmin focused on lending and credit structuring during his investment banking career and launched Indonesian peer-to-peer lender GetMo in 2023.

Running GetMo exposed him to industry constraints like limited underwriting capacity. He pivoted to found Moneta to build tools that address these pain points.

😟 Problem

Lenders across Southeast Asia face severe capacity constraints that restrict lending to small businesses. A recent peer-to-peer lending collapse flooded banks with quadruple the application volume compared to 2023.

Only 28% of borrower files reach a decision within seven days of data submission and manual reviewers take up to 14 days to process applications. This means banks lose revenue as qualified borrowers get stuck in the pipeline.

💡 Solution

Moneta provides an AI credit underwriter that accelerates the decision-making process for financial institutions. The platform handles manual document analysis and offers several core capabilities. It also offers:

• Real-time analysis of unstructured documents like bank statements and invoices.
• Automated generation of end-to-end credit memos for committee review.
• Decision turnarounds reduced from weeks to minutes.
• Average client cost reduction of 83%.

Image credit: Moneta

📊 Market size

Southeast Asia has an estimated small-medium-sized enterprise (SME) credit gap exceeding US$300 billion. The company targets over 1,500 lenders in the region, starting in Indonesia.

Projected annual recurring revenue for the Indonesian market totals US$20 million. Regional projections scale to US$95 million.

🤝 Team

• Sergei Kuzmin. Co-founder and CEO. Previously scaled a credit fintech in Indonesia to US$1 million ARR in seven months.
• Denis Anoshin. Co-founder and CTO. Orchestrated proprietary AI models in international banks since 2007.
• Kartik Mandar. Head of machine learning. Built machine learning methods for astrophysics research in India and adapted models for corporate fintech.
• Nando Hidayat. Chief intelligence officer. Former Indonesian country lead at competitor Perfios.

🚀 Traction

• Secured 55 paying customers.
• Reached US$54,000 in monthly recurring revenue in May.
• Reduced average bad loan rates by 2% across its client base.
• Maintained an 11-day average sales cycle from initial meeting to payment.

🏆 Competition

💰 Financials

🚩 Risks