Conversus Prebuilt Models
All of Converseon’s prebuilt LLM-based models were built for one core purpose: to help insights professionals more efficiently and effectively tackle key insights use cases.
Converseon’s prebuilt models can be purchased and deployed in special packages that are aligned to these use cases, from basic voice segmentation and sentiment analysis to customer experience analysis, brand tracking, reputation measurement, and more.
Sentiment
Converseon’s LLM-based sentiment model provides best-in-class sentiment analysis for social listening, voice-of-customer, earned media, and other unstructured text data.
With industry and use case-specific fine-tuning and the ability to generalize to new topics/domains, Converseon’s sentiment model routinely achieves accuracy in the 85-95% range.
Intensity
Expressions of opinion are not binary. How strongly an opinion is expressed is critically important.
Our LLM-powered “Intensity” model enables you to understand the strength of consumer opinions helping to identify what opinion may result in action and impact. Key use cases include customer experience (such as alignment with NPS), brand health, crisis communications and more.
Content Type
Social listening data ≠ voice-of-consumer data.
Getting to real stakeholder voices in noisy social media conversations is challenging. Unstructured datasets often intermingle VoC content with noise/bots, ads/spon-con, owned content and earned media.
The Content Type model automates the voice segmentation that is central to these intelligence efforts, accurately segmenting unstructured conversation data by News, Promotional, and User-Generated Content.
Consumer Attitude
Consumer Attitude is the essential model for analyzing consumer mindset and CX in unstructured conversation data.
Where traditional Emotion analysis often lacks accuracy and actionability, our LLM-based Consumer Attitude model provides a highly accurate and nuanced view of consumer relationships to brands/products.
The model teases apart 15 different attitudes (often within a single record), ranging from Aggrieved or Outraged, to Cautionary or Inquiring, to Grateful or Laudatory.
Political Landscape & Brand Safety
How engaged with political issues are the people mentioning your brand online? How do these political and public affairs issues impact your brand, if at all?
In our polarized political climate, understanding the impact of these issues is more important than ever for maintaining brand safety. Our LLM-based model will accurately identify politically charged conversations and help you quickly identify and stay ahead of opportunity and risk areas for your brand perception and reputation.
Environment, Social & Governance (ESG)
Our LLM-based ESG model provides comprehensive and detailed classification of your brand and competitors across Environment, Social and Governance as well as more than 30 sub-topic classifications ranging from Carbon Offsets to Labor Rights.
This model is a cornerstone of reputation measurement and ESG impact analysis in unstructured news and social media data. Our measurements of brand perception surrounding ESG issues have proven to be predictive of concrete business outcomes such as revenue, share price and sales.
Operational Efficiency and Cost Savings
