After suspected AI-driven attacks exposed customer data at Shinhan, KB Kookmin, Hana and other banks, South Korea's Financial Services Commission ordered firms on 2 October to inventory external systems and close authentication gaps.
Gwanak Lab raises 300 million won from Seoul Techno Holdings for financial decision AI
EditorialBy TrustList Editorial
Seoul-based Gwanak Lab, which builds a decision engine that helps banks, card companies and insurers act on customer data, has raised 300 million won (about USD 228,000) from Seoul Techno Holdings, WOWTALE reported on 2 October 2026.
- South Korea
- Seoul, South Korea
- Banking Software
- Predictive Analytics
- +2 more
About Gwanak Lab raises 300 million won from Seoul Techno Holdings for financial decision AI
Gwanak Lab raises 300 million won from Seoul Techno Holdings for financial decision AI
2 October 2026: Gwanak Lab, a South Korean startup that builds AI to support decisions in financial institutions, has raised 300 million won (about USD 228,000) from Seoul Techno Holdings, according to a WOWTALE report of 2 October 2026. The investor is the technology holding company linked to Seoul National University.
Not yet independently verified. No announcement from the company or its lead investor has been found; this round is reported only by the press, and the amount and investors are the outlets' figures. WOWTALE says the company announced it on 2 October, but we have not found that announcement. The stage of the round is not stated. We will update this when it can be confirmed, and remove this note.
The round
According to WOWTALE, Gwanak Lab announced the investment on 2 October. The report names Seoul Techno Holdings as the only investor and does not state a stage. The company plans to use the money to advance its decision engine, grow its development team, and strengthen its service infrastructure to meet financial institutions' requirements for security, data processing and network-isolated environments.
What the company does
Founded in 2023, Gwanak Lab builds tools for tasks such as preventing customer churn, prioritising debt collection and screening policy loans, for banks, card companies, insurers and public finance bodies. Its engine combines structured data, such as transaction history and credit information, with unstructured data, such as customer-service records and transaction descriptions, to build a single view of each customer. It forecasts how a customer's situation may change, presenting ranges rather than single estimates, and suggests which customers need attention and what action to take, with the reasoning shown. The company is running proof-of-concept projects with financial institutions and launched a consumer investment service on the same engine in July.
Why it matters for buyers
Korean financial regulators expect explainable decisions and strict data isolation, and many institutions cannot send customer data to outside clouds. A vendor that builds for network-isolated deployment and shows its reasoning addresses those constraints directly. It is also very early: a small round and pilots rather than contracts.
What to check
- Whether the engine can run fully inside your network, and what data leaves it.
- How its reasoning is presented and whether it satisfies your model-risk and fair-lending reviews.
- How it handles unstructured data such as call notes, including personal information.
- Results from pilots, measured against your current models.
- The company's capacity to support a production deployment at its current size.
Company profile on TrustList: Gwanak Lab
Related on TrustList:
Sources
Categories & features
- South Korea
- Seoul, South Korea
- Banking Software
- Predictive Analytics
- Debt Collection Software
- Venture Capital
TrustList Weekly
The week in software and IT, in one email
The news that matters to buyers, new rankings and our own research. Every Thursday, free, and easy to leave.
You may be interested in
The first penalties under Korea's 2021 in-app payment law also order both companies to allow payment outside the app; Apple is found to have charged Korean developers commission on VAT-inclusive prices for seven years.
Someone used credentials for Amadeus system accounts between August 2023 and September 2026. Asiana says booking reference numbers were taken and that Amadeus is still working out what else was.
South Korea's science ministry said on 2 October that its sovereign AI foundation model project continues, with SK Telecom, LG AI Research and Upstage competing for two places, while a new frontier-model programme is proposed in the 2027 budget.
Two K-Startup calls close this week: KISED's Silicon Valley go-to-market support for Korean startups (9 October, 18:00 KST) and the 2027 Super-Gap Startup Project call for host organisations (12 October, 15:00 KST).
From 1 November 2026 Pakistani banks and DFIs take SME financing applications only through the State Bank's new portal, which shows applicants their status and rejection reasons and gives the regulator a live view of every lender.
The Information Regulator also found the bank committed a criminal offence under POPIA, and fined a Johannesburg college R500,000 for ignoring an enforcement notice it had 120 days to meet.
Pakistan Express Clearing System runs through the NIFT clearinghouse, with an extra daily clearing cycle in the pilot. Every member bank has to be ready for nationwide coverage by 30 April 2027.
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- More South KoreaThe ranking for this subject
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- AI ModelsBenchmark scores and community ratings for every major model.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.