NH Investment Securities Plans Firmwide AI Platform

NH Investment Securities has selected a provider for its firmwide AI platform and is preparing to begin the project, with construction targeted for the end of this year. The company plans to use the platform to redesign core workflows and connect a range of AI models with internal data and business systems. AI work is handled by the AX Promotion Department under the Management Strategy Division, which has fewer than 10 employees. An existing chatbot uses retrieval-augmented generation (RAG, a method that retrieves relevant information before generating answers) to answer employee questions using internal manuals and information as a knowledge base.

The planned platform will let employees build work-focused AI agents in no-code and low-code environments, then connect the agents' task steps into agentic workflows. The company says this will expand AI use led by business departments. During development, it is reviewing procedures, required data, and final output formats with employees who perform the work. NH Investment Securities says it is designing workflows to distinguish tasks handled by AI from decisions and responsibilities retained by people, while linking multistep tasks into a consistent process.

The company is also building an enterprise data lakehouse, which integrates and manages varied data. The aim is to reduce repeated effort in gathering and processing materials for different tasks and to connect reliable data to AI securely. After establishing the platform this year, the company aims to introduce agents gradually into key processes across business divisions starting next year to improve profitability. It also plans to spread a bottom-up approach in which employees develop and improve agents themselves.

To improve answer accuracy, NH Investment Securities plans to introduce harnesses that manage answer generation, LLM judges—separate models that evaluate AI responses—and human-in-the-loop review, in which people check results. It is preparing governance for the full AI service lifecycle, from planning and development through deployment and operation. The company plans to refine risk assessments, roles and responsibilities, approvals, and operational checks with relevant departments, and connect these procedures to existing information security, compliance, and IT management systems. Its approach will reflect the AI Basic Act and financial-sector AI guidelines, and cover personal data protection, security, and internal controls.

Starting next year, NH Investment Securities plans to expand workflow redesign and agent deployment in stages, set performance indicators for each agent, and assess the results. It expects to allocate resources to areas with greater impact and increase its AI investment budget. The company identifies its firmwide platform and governance, together with support for employee-led agent development, as competitive strengths, and expects AI to help improve profitability in securities firms' core business operations.

NH Investment Securities has selected a provider for its firmwide AI platform and is preparing to begin the project, with construction targeted for the end of this year. The company plans to use the platform to redesign core workflows and connect a range of AI models with internal data and business systems. AI work is handled by the AX Promotion Department under the Management Strategy Division, which has fewer than 10 employees. An existing chatbot uses retrieval-augmented generation (RAG, a method that retrieves relevant information before generating answers) to answer employee questions using internal manuals and information as a knowledge base.

The planned platform will let employees build work-focused AI agents in no-code and low-code environments, then connect the agents' task steps into agentic workflows. The company says this will expand AI use led by business departments. During development, it is reviewing procedures, required data, and final output formats with employees who perform the work. NH Investment Securities says it is designing workflows to distinguish tasks handled by AI from decisions and responsibilities retained by people, while linking multistep tasks into a consistent process.

The company is also building an enterprise data lakehouse, which integrates and manages varied data. The aim is to reduce repeated effort in gathering and processing materials for different tasks and to connect reliable data to AI securely. After establishing the platform this year, the company aims to introduce agents gradually into key processes across business divisions starting next year to improve profitability. It also plans to spread a bottom-up approach in which employees develop and improve agents themselves.

To improve answer accuracy, NH Investment Securities plans to introduce harnesses that manage answer generation, LLM judges—separate models that evaluate AI responses—and human-in-the-loop review, in which people check results. It is preparing governance for the full AI service lifecycle, from planning and development through deployment and operation. The company plans to refine risk assessments, roles and responsibilities, approvals, and operational checks with relevant departments, and connect these procedures to existing information security, compliance, and IT management systems. Its approach will reflect the AI Basic Act and financial-sector AI guidelines, and cover personal data protection, security, and internal controls.

Starting next year, NH Investment Securities plans to expand workflow redesign and agent deployment in stages, set performance indicators for each agent, and assess the results. It expects to allocate resources to areas with greater impact and increase its AI investment budget. The company identifies its firmwide platform and governance, together with support for employee-led agent development, as competitive strengths, and expects AI to help improve profitability in securities firms' core business operations.