THE CUSTOMER
The client offers multinational financial services. Their headquarter is in Washington DC. The client funds social development projects in Latin America and the Caribbeans by providing both financial and technical support. The Bank mainly focuses on providing solutions for social incorporation and equality, increase the productivity and add innovation and economic integration. They mainly aim to address issues such as gender equality, diversity, climate change, environmental sustainability, institutional capacity, and the rule of law.
BUSINESS GOALS
The bank assists 100+ offices sprawled across the globe and supports a multilingual workforce. The bank’s officers have worked on various projects and gained a lot of real-world experiences. They have a variety of skills, lot of operational knowledge and have developed themselves as the Subject Matter Experts over certain key topics areas. These officers work from geographically spread locations using documents that are in English and/or Spanish and/or Portuguese or other languages.
The primary business goal and use case is to create a single integrated knowledge management platform for individuals to access, identify and reach out to individual SME’s seeking help on current projects and tasks at hand.
THE SOLUTION
ThirdEye worked with the client to implement a cognitive computing application leveraging IBM Watson services on the IBM Bluemix cloud infrastructure. The overall solution included a virtual agent that understands natural language query inputs and uses cognitive deductions to respond to the conversation; simulating a human-like conversation between users and the system.
ThirdEye leveraged AlchemyAPIs for advanced text analytics including keyword extraction, entity extraction, sentiment analysis, emotion analysis, concept tagging, relation extraction, taxonomy classification, author/person extraction and relevance score extraction. The output was further processed to eventually create JSON structures with all relevant metadata and associated information of every SME.
SMEs were retrieved and ranked in real time as per the user’s queries
The process steps were as follows:
- Process data from multiple sources including
online publications, project documentations, blog
postings, employee information from internal
systems and online social media platforms. - Extract entities and metadata information about
each entity. - Analyze all information processed so far to
identify “Subject Matter Experts” on various topics
across the enterprise. - Rank the SMEs, based on contributions,
publications, etc. and create their deep profiles. - Enable users to query using simple English
sentences through a “chat” interface to find &
connect with relevant subject matter experts. - Perform drill down analysis on each ranked SME
and look up all associated information & facts
Technologies Incorporated:
- IBM Watson Developer Cloud
- Alchemy API – for text analysis through natural language
processing. - Document Conversion – for converting various
document formats to text documents. - Language Translator – to support multilingual
conversations with users. - CloudantDB – for storing the output from AlchemyAPI
and other unstructured data. - DashDB – for storing structured data and analytics.
Watson Conversations – for automating interactions
with end users. - Insights for Twitter – to load and analyze Twitter data.
Suite of custom Java Applications for crawling, extracting
and processing data from multiple sources - Purpose built search and ranking algorithms for SME
search and ranking. - Custom NodeJS chat bot interface
VALUE CREATED
The bank now believes that productivity across the globe will significantly increase due to easy identification of Subject Matter Experts. Such enhancements will improve operational efficiencies by a margin of 30% to 40% on all funding and technical evaluations. Additionally, as the scope of the system further expands to include multiple uses cases including but not limited to knowledge management of subject expertise, cross-border collaboration and employee engagement across project implementations will improve.