How Italgas has harnessed generative AI for customer complaint management
Image: Databricks
Italgas has implemented a generative AI (GenAI) solution to better manage and resolve customer complaints.
The inhouse developed solution by Bludigit, Italgas’ digital company, was aimed to transition the process of handling customer complaints – numbering over 90,000 in 2023 – from a manual process involving multiple systems and databases to a more automated system enabled through digitalisation.
Previously with a complaint the operator would be switching between multiple systems, including Salesforce for complaint management, meter data management for usage data and its GasToGo for technical information.
At the heart of the new system is the Databricks Lakehouse platform, which serves as the central data repository and provides operators a comprehensive view of each case.
Have you read?
AI in Europe’s energy sector – working towards a common understanding
How genAI can help accelerate the energy transition
This can include, for example, real-time meter readings, current alarm statuses and even future inspection schedules for the meter in question.
The solution is based on the retrieval augmented generation (RAG) AI technology that learns from past successes.
When a new complaint arrives, the system analyses the complaint details and searches through the historical database to identify similar past situations and extract successful resolution patterns.
Combining these with current technical data, a comprehensive roadmap for resolution is generated in the form of detailed instruction notes to guide operators through the resolution process.
In the final stage of resolution operators can provide immediate feedback on the suggestions’ usefulness, which in turn helps the system learn and adapt and thereby constantly improve the quality of its recommendations.
According to a case study blog post co-authored by Bludigit, the implementation of this GenAI solution has transformed its complaint resolution process.
It has delivered reduced handling times through automated data gathering, provided more consistent resolutions through AI-powered recommendations, enhanced operator efficiency by eliminating manual data collection and improved customer satisfaction through faster, more accurate responses.
Looking ahead plans include expanding the system to handle additional complaint types and developing autonomous resolution capabilities for straightforward cases.
In addition the RAG implementation is expected to be further enhanced with more sophisticated retrieval strategies.
