Treasurer Brings On-Premises AI to Financial Institutions

via ACCESS Newswire
ⓘ This article is third-party content and does not represent the views of this site. We make no guarantees regarding its accuracy or completeness.

SINGAPORE, SG / ACCESS Newswire / November 16, 2026 / Treasurer Pte. Ltd., a financial AI company established in Singapore in 2023, helps financial institutions and hedge funds adopt AI across their daily work. Treasurer provides an on-premises AI system and a finance-specialized large language model (LLM) that run on hardware the client owns.

Financial institutions face three hurdles with AI. Analysts lose time gathering data, matching figures and building tables. Client data and internal reports often cannot be sent to outside AI services. Compliance teams need records of the source behind each figure and who reviewed each result. Treasurer handles this work inside the institution's own office.

Treasurer's work with each client starts with one existing workflow, such as research, due diligence or compliance review. Treasurer maps the steps, sets up an AI workspace around them, installs and connects the system, and assigns a dedicated engineer. A four-week pilot includes a parallel run alongside current work. It is evaluated on time per case and source accuracy before any wider rollout.

The system is built in four layers:

  • Workspace: Treasurer configures the workspace around the client's existing workflow. AI handles each step, and a staff member reviews it.
  • Financial data engine: Treasurer's AlphaLenz engine collects and cleans filings, market data, news and market rumors from more than 6,800 sources in five countries. It links figures to their original sources. Data flows one way into the office; documents, questions and answers are not sent to Treasurer.
  • Finance-specialized LLM: Treasurer's own TA7Z model, trained for financial work, runs on the client's hardware. On supported hardware, a router keeps sensitive requests, such as client data or positions, on TA7Z inside the office. General questions go to an outside model only if the client allows it.
  • On-premises: the system runs on hardware inside the client's office or on the client's own GPU servers. Clients choose and own the machine; Treasurer does not sell hardware.

In practice, one question returns a finished report within minutes, built from the firm's own research, notes, filings and data, with the source sentence behind each number. The system reads Korean corporate filings in full and works in Korean, English, Chinese and Japanese. It supports research and review; decisions stay with the client's team.

For IT and compliance teams, each client runs a separate, single-tenant environment with encryption keys held by the client. Every answer is logged with its model, sources, processing location and reviewer, and an inspection pack downloads in one step for regulatory reviews. Treasurer is ISO/IEC 27001 certified.

"Financial institutions want to use AI, but much of their most valuable data cannot leave the building," said Ted Kim, CEO of Treasurer. "We place the system and the model inside the client's office, so their teams can put the data they already have to work and see the source behind every answer."

Request a 20-minute demo at treasurer.net/core.

About Treasurer

Treasurer Pte. Ltd. is a financial AI company established in Singapore in 2023, with its office at 16 Raffles Quay, Hong Leong Building. It helps financial institutions and hedge funds adopt AI, building on its proprietary financial ontology and its finance-specialized TA7Z model. Its products include AlphaLenz, an investment analysis platform. Treasurer is a member of NVIDIA Inception and is ISO/IEC 27001 certified.

Media Contact

Treasurer Pte. Ltd.
Email: ryan@treasurer.co.kr
Website: treasurer.net

SOURCE: Treasurer Co., Ltd.



View the original press release on ACCESS Newswire

Report this content

If you believe this article contains misleading, harmful, or spam content, please let us know.

Report this article