Midas Analytics Uses AI to Solve Information Fragmentation Across Asian Markets【Hong Kong ICT Awards 2025: Startup Award】

When conducting market research or making investment decisions in Asia, language barriers and fragmented information are often the biggest pain points. Regulatory filings, news, and stock exchange announcements are scattered across jurisdictions and are frequently published only in local languages. Winner of the Bronze Award in the “Hong Kong ICT Awards 2025: ICT Startup (Software & Apps) Award,” Midas Analytics is an AI-driven market research platform built specifically for Asia and designed to turn messy, dispersed data into auditable decision intelligence.

After receiving the award, Midas Analytics founder Dr. Michele De Filippo shared the team’s excitement and vision:

“We’re thrilled and deeply honored to be among Hong Kong Best SaaS webapps of 2025. It’s a strong validation of our mission to turn fragmented, multilingual market information into auditable, actionable intelligence. We’re grateful to the judges and our customers for the trust and momentum to keep building.”

________________________________________________________________________________

Built for Asia: A Dual Engine of AI Agents and AI Assistants

Unlike the US and European markets, where data is more centralized, critical signals in Asia are often dispersed across regulators, exchanges, trade publications, and niche local websites. Midas Analytics has a clear positioning: an “Asia-first” AI research platform.

Dr. De Filippo explained that the platform is powered by two core engines:

• AI Agents: Continuously tracking companies, industries, or topics. When material changes occur, the agents automatically generates structured briefings and alerts, hence removing the burden of manual monitoring.
• AI Assistants: Users can query a curated database directly, and the AI returns decision-grade answers within seconds. Crucially, every answer includes inline citations and original excerpts, ensuring the content is verifiable and compliance-ready.

By combining multilingual data ingestion with rigorous source governance, Midas Analytics helps financial institutions and corporate strategy teams significantly reduce the time and risk associated with cross-border research.

MIDAS AI vs. Traditional LLMs: A Different Approach to Research

While many are familiar with general-purpose Large Language Models (LLMs) like ChatGPT and Gemini, MIDAS AI is engineered for a different purpose. Dr. De Filippo elaborated on the distinction:

“We are not aiming to compete with traditional LLMs; in fact, we leverage their power in our platform. The key difference lies in our specialized approach. Traditional LLMs are designed to search vast, public datasets, which is powerful for general knowledge but can be a limitation for specialized research. MIDAS AI, on the other hand, is built on a curated and pre-indexed database of micro-research and niche information. This allows our users to first select where to look for answers, giving them control over the source and quality of information.”

This specialized database, combined with the platform’s architecture, enables MIDAS AI to deliver deeper insights at a much faster pace than generalist models. “Because our data is already structured and indexed for financial and market research, we can uncover connections and provide a level of detail that would be difficult to achieve with a general-purpose LLM,” Dr. De Filippo added.

He concluded with an open invitation: “We encourage everyone to try it out for themselves. The best way to understand the difference is to experience it firsthand. We are confident that you will see the difference yourself.”

The pain point: “information lag” in multilingual, unstructured data

Dr. De Filippo noted that the product was born out of a long-standing workflow gap in Asia-focused research: “The most valuable signals are often released first in local languages, and in formats that are hard to monitor at scale.” This creates a persistent “information lag” for investment and corporate strategy teams because analysts must manually piece together scattered announcements and news, often discovering key updates too late.

To compress this cycle, the team overcame three major technical challenges:

• Multilingual retrieval and normalization: Asian markets span multiple writing systems and domain-specific terminology, requiring strong language processing capabilities.
• Source heterogeneity: Key disclosures come in many forms (PDFs, scanned documents, HTML tables, and more) each demanding different extraction techniques.
• Auditability: Generating answers is not the hard part; generating answers that stand up to scrutiny is. The team built a strict citation framework to ensure every AI-generated claim is traceable to evidence.

What’s next: Deepening “Asia-first” and improving signal-to-noise

Looking ahead, Midas Analytics plans to further deepen its “Asia-first” moat by expanding coverage of additional Asian markets and local-language sources. The team will also work to improve AI precision to reduce irrelevant noise, while strengthening enterprise-grade capabilities such as permission controls and audit trails.

From a go-to-market perspective, the company will focus on sectors most burdened by fragmented information (like financial institutions and corporate strategy teams). Longer term, Midas Analytics aims to replicate its “multilingual + localized sources” model in other cross-border regions facing similar research challenges.

From “gateway” to “global builder”: Hong Kong’s next phase

On the future of Hong Kong’s innovation ecosystem, Dr. De Filippo argued that the city has strong fundamentals, while stressing that the next stage will hinge on execution:

“Hong Kong is well positioned to be a high-impact I&T hub by combining world-class capital markets, strong universities, and a credible legal and regulatory framework with closer integration into the Greater Bay Area. The next phase will be defined by execution: faster commercialization of research, deeper industry–startup collaboration, and more cross-border talent and data flows especially in AI & Fintech. If Hong Kong continues to reduce friction for founders and scales pathways from pilot projects to procurement, its ecosystem can shift from ‘gateway’ to ‘global builder’.”