Siemens, the multinational industrial manufacturing conglomerate, innovates with cutting-edge technology.
Siemens uses Microsoft Viva Engage (formerly Yammer), the Microsoft 365 enterprise social networking service, as its corporate platform for internal communication. The company sought to enhance the collaboration platform with generative artificial intelligence (AI).
Key Challenges
- Inability to prioritize tax-related updates on Microsoft Viva Engage (formerly Yammer)
- Inability to generate actionable insights from social posts
Ready to experience?
TALK TO EXPERTS
The Solution
Generative AI model helps tax consultants focus on social posts related to taxation, including updates and regulatory changes
Siemens partnered with Infosys to implement a generative AI solution to classify and summarize tax-related communication on Microsoft Viva Engage (formerly Yammer).
A small set of labeled social posts was readily available, but it was inadequate to train a conventional machine learning model. Errors due to manual classification of social posts in multiple languages added to the complexity.
Infosys combined its consulting expertise, with applied AI capabilities - part of Infosys Topaz to develop a large language model (LLM)-based solution. Our GPT specialists created in-context prompting workflows for classifying conversations and generating summaries.
Automated Classification Ensures Accurate Labelling of Unstructured Data
- ~ 90% accuracy in binary classification
- 85%-90% accuracy for multi-class classification
- < 5 seconds to classify each post
Benefits
Accurate classification of social conversations based on category
Automated summarization of social media posts
