Democratising access to data with a virtual assistant
Target Reply has partnered with Agos to create a virtual assistant that simplifies interaction with data and supports the transformation of the Data Warehouse.
,allowExpansion)
SCENARIO
Simplifying access to data in a complex context
Agos, among the main operators in the consumer credit sector, has always focused on the analysis and management of data to support strategic decisions. To improve the autonomy of business users and increase the efficiency of analytical processes, speeding up the search for information, the company has launched a program for simplification and democratisation of data.
SOLUTION
A virtual assistant to facilitate interaction with data
In this context, a solution based on a AI assistant has been designed and implemented with the support of Target Reply, capable of answering questions about the origin of data, the meaning of fields, and the relationships between tables, allowing for independent analysis activities. The solution is based on the integration of LLM models and architecture RAG (Retrieval-Augmented Generation), which allows for real-time consultation of an updated corporate knowledge base and provides contextualised answers.
The system is also able to ask the user for clarifications to improve the relevance of the responses. The assistant behaves like a virtual data steward, supporting the user in the correct use of information, providing clarifications on any limitations of data access. The same solution has been extended to support the project of reviewing business systems, which also involves a redesign of the Data Warehouse, where the assistant has been used to facilitate the search for information within the new data sources and for comparing the current data models with those in evolution.
How it was made
To develop the solution, a multi-phase process was defined, starting from the analysis of business content to the implementation on RAG architecture. The knowledge base was built by integrating technical documentation, business definitions, FAQs, and historical tickets, then indexed within a vector database to enable semantic search.
The AI assistant has been trained to understand user requests, select the most relevant sources, and formulate contextualised responses. The system maintains the state of the conversation, allowing for natural and continuous interactions.
In case of ambiguity or generic requests, the assistant is able to ask clarifying questions to guide the user towards the desired information.
The solution was developed iteratively, with cycles of testing and continuous improvement, directly involving end users to optimise language, behaviours, and information coverage.
Simplified access
to the data also for non-technical users
Reduction of time
for analysis and assistance thanks to the autonomy enabled by the assistant
Support for the renewal of the DWH with automated matching activities between data sources and documentation
Agos is one of the leading Italian companies in the consumer credit sector. With extensive experience in providing personalised financial solutions, Agos offers innovative and accessible products to individuals and families, including personal loans, credit cards, and financing solutions for the purchase of goods and services. The company is known for its customer focus, transparency of conditions, and commitment to promoting responsible credit management.
Target Reply is the company of the Reply group specialized in creating Big Data and Advanced Analytics solutions. Target Reply supports companies from identifying needs to designing and implementing solutions through data integration technologies, data modeling, predictive analysis, using the most innovative tools in the field of Business Discovery and Big Data. Target Reply has gained significant experience with major Italian and foreign corporate groups and is capable of operating in all major markets: Telco, Finance, and Manufacturing.