Agent Watson

Accelerating Searches, Empowering Investigators

Agent Watson Logo

Introducing Agent Watson

Agent Watson is a web-based platform that leverages a multi-agent open web information gathering system to automate early-stage digital information gathering and support decision-making in missing persons investigations. It performs targeted data retrieval, multimodal analysis and iterative relevance assessment to enable faster and more consistent intelligence gathering during the critical initial stages of an investigation.

Team members

Jagannadha Rao Surya Vijapurapu (ESD), Gay Kai Feng Matthew (DAI), Ernest Tan Wei Yan (ISTD), Alaguvignesh Thirunavirkarasu (ESD), Cheryl Kwek Tze Theng (ISTD), Chia Tang Hsieh (ISTD)

Instructors:

  • Dorien Herremans

Writing Instructors:

  • Susan Wong

What is Agent Watson?

Agent Watson is a web-based platform that supports early-stage missing persons investigations using a multi-agent open web information gathering pipeline. An orchestrator coordinates specialised agents—Leader, Researcher, Processor, and Judge—to iteratively gather, analyse, and refine information, transforming unstructured online data into actionable insights. The system leverages a modular MCP architecture for flexible tool integration and is accessed through a React-based web portal that provides clear reports and visualisations for faster decision-making.

Problem Statement

How might we accelerate the missing person search by automating the collection and analysis of publicly available information online to support decision-making?

 

A reliable, repeatable, and explainable system for automating early-stage open web information gathering collection and assessment is therefore essential to address the growing volume and diversity of online content as manual workflows are becoming increasingly unsustainable. 

Existing Solutions and their Limitations

Existing solutions such as public AI search engines and commercial open web information gathering platforms provide partial support but present key limitations. These include high noise levels, limited verification capabilities, lack of transparency, privacy concerns, and reliance on proprietary systems.
/

Key Features

Investigator Web Portal
Multi-Agent Pipeline
Modular Tool Integration

A React-based web interface provides investigators with an accessible way to interact with the system. Users can submit search queries, monitor investigation progress, and explore results through interactive visualisations, knowledge graphs, and structured reports. This interface ensures that investigators can quickly interpret the intelligence gathered and make informed decisions during time-critical missing persons investigations.

 

At the core of the system is a multi-agent AI pipeline coordinated by an orchestrator. Each agent specialises in a specific task within the investigation workflow:

  • Leader Agent – Develops the overall investigation strategy and identifies key information gaps.
  • Researcher Agent – Collects raw data from public online sources using open web information gathering tools.
  • Processor Agent – Extracts and structures relevant information from collected data using techniques such as entity recognition and relationship analysis.
  • Judge Agent – Evaluates whether sufficient information has been gathered or if further searches are required.

These agents operate in an iterative loop, continuously refining the search until meaningful intelligence is obtained. Once complete, the system generates a concise summary report and a detailed evidence report with source references.

The system integrates investigation tools through Model Context Protocol (MCP) servers, allowing AI agents to access specialised capabilities such as web scraping, data extraction, and entity analysis. This modular design enables tools to be added, replaced, or upgraded without modifying the core agent logic, ensuring long-term maintainability and flexibility.

Key Contributions of Agent Watson

  • Solving the problem of Data Explosion (Webscraping):

    Agent Watson replaces broad, noisy scraping with targeted extraction and automated filtering, ensuring investigators receive high-quality signals instead of a massive volume of irrelevant data.

  • Eliminating AI Hallucinations (Public LLMs):

    By using grounded reasoning and transparent workflows, Agent Watson ensures every AI-generated insight is cross-referenced with scraped facts, removing the risk of fabricated information.

  • Extensible High-Signal open web information gatherer (Aggregators):

    Unlike rigid, proprietary aggregators, Agent Watson’s modular Python-based architecture allows users to easily customize or add new investigative modules, providing high-signal results without black box limitations.

In partnership with :

Acknowledgements

Team S24 would like to thank our Capstone instructors: Dr Dorien Herremans and Dr Susan Wong, for their insightful advice which were pivotal to Project Watson’s success. 

The team would also like to thank our mentors from the Home Team Science and Technology Agency: Dr Terence Tan, Mr Arka Ray and Mr Tsay En Zhan, for their valuable guidance and support throughout the project.

Menu

ornament-menu
Read More
Read More
Project E10: PactPro - Automated Mergers & Acquisitions
Project E10: PactPro – Automated Mergers & Acquisitions PactPro is...
Read More
Proj E06 - Interstellar - GoldNest
CuliNexus The Digital Backbone of small-and-medium enterprises (SME) Kitchens A smart culinary platform...
Read More
Project E05 - Doby
Problem Context We spoke to over 30 SME owners and operators across Singapore to understand the problem’s...
Read More
Project E02 ATHANOR by BUNGUS
How does Athanor work? Athanor is a plug-and-play fermentation monitoring system that gives breweries...
Read More
Proj 11 - Delta - Smart Meeting Pods
PROBLEM Hybrid and remote work is becoming the new norm, especially as...
Read More
Parents Lah!
A mobile-first, scenario-based casual game grounded in Self-Determination Theory and Baumrind’s...
Read More
Proj 39 - Esplanade - Design Accessibility
Solution Renders and Prototype This inclusive design at Esplanade combines an interactive tactile environment...
Read More
Proj S23 - SITA - Autonomous Luggage Concierge
Problem Air travel has become more accessible than ever. However, for passengers with restricted mobility,...
Read More

Contact the Capstone Office :

+65 6499 4076

8 Somapah Road Singapore 487372

www.sutd.edu.sg/Capstone

capstone@sutd.edu.sg

About Capstone Design Programme

At Singapore University of Technology and Design (SUTD), we believe that the power of design roots from the understanding of human experiences and needs, to create for innovation that enhances and transforms the way we live. This is why we develop a multi-disciplinary curriculum delivered v ia a hands-on, collaborative learning pedagogy and environment that concludes in a Capstone project.

The Capstone project is a collaboration between companies and senior-year students. Students of different majors come together to work in teams and contribute their technology and design expertise to solve real-world challenges faced by companies. The Capstone project will culminate with a design showcase, unveiling the innovative solutions from the graduating cohort.

The Capstone Design Showcase is held annually to celebrate the success of our graduating students and their enthralling multi-disciplinary projects they have developed.

/