Main Challenges for BFSI in Market Monitoring:
1. The surge in social media usage presents notable challenges in extracting insights from the vast and diverse data sources.
2. The manual-intensive process of tracking and analysing all sources is costly and unsustainable, e.g:
- Regulators sifting through social platforms to detect financial irregularities like mis-selling and market manipulation.
- Compliance teams monitoring client-related news and social references impacting business decisions.
3. The existing SaaS-based products on the market do not meet the high security standards required for compliance and lack the flexibility needed to customise workflows for financial institutions. Additionally, they fail to cover the full range of data sources necessary for the financial sector.
4. Conventional approaches, such as Interactive Voice Response (IVR) systems and customer satisfaction surveys, often fall short in capturing the immediacy and nuance of customer feedback needed to identify market risks, which may lead to a crisis at the end.
Solution highlights:
Advanced NLP Algorithms
Utilise state-of-the-art NLP algorithms, LLM models and leverage proprietary technologies, including smart search engine, RAG, chatbot, along with self-improving machine learning algorithms.
Data Consolidation
Facilitate data collection and ingestion through data crawling and API integration, gathering relevant information from social media, news sources, and various external/internal sources into a single platform.
Fully Customisable Solutions
Offer customisation features to all components tailoring to the unique workflows and requirements for comprehensive text analysis.
On-premises Deployment
Address the concerns of BFSI institutions regarding data privacy, security, and the control of sensitive information by offering an on-premises deployment.
Trilingual Support
Analyse text in multiple languages, including Traditional Chinese, Simplified Chinese, English, Cantonese slang, idioms, and buzzwords, as well as mixed-language dialogues. PRISMA can comprehend and extract insights from internet slang and colloquial expressions, which is unique in the region.
Business Intelligence Driven
Offer user-friendly visualisation tools, interactive dashboards and automate key case management processes, including alert generation, empowering users with real-time analytics to proactively identify and respond to emerging risks, compliance issues, hidden insights and market trends.
Main Functions:
- Self-served model training
- Taxonomy construction
- Conceptual intelligent search
- Sentiment analysis
- Content categorisation with relationship visualisation
- Unstructured and structured data comprehension
Case Study
Financial Institute
Identify critical elements that affect stock price
The Institute wanted to filter unimportant information in newsletters in order to identify critical elements that might affect stock price.
Public Securities Bureau
Speed up intelligence handing to prevent crimes
Faced with a large amount of data across a wide range of formats, the Bureau needed a system that could improve the ability and quality of handling intelligence, to better prevent crime.
Public Transportation Corporation
Develop customer loyalty with shorter time-to-insights
Operating one of the largest railway systems in China, the public transport service provider receives a large amount of daily passenger feedback. A system that could automatically and efficiently categorise these texts for analysis would significantly improve operations.
Non-governmental Organisation
Analyse the young generation on social media
Better understand the attitudes of young people who use social media towards drug problems, and grasp the current trends and correlations of various drug problems on social media.