Twig

The platform ensures AI remains current by refreshing data from private sources weekly or at configured intervals. This process keeps responses accurate and aligned with the latest information. Personally identifiable information is automatically filtered to protect confidentiality while maintaining high-quality training.

Twig software allows enterprise teams to refine and adjust AI outputs according to internal knowledge and nuances. Users can define persona settings, including tone and preferred data sources for responses. Fine-grained controls also let teams regenerate answers, adding or removing data sources to ensure outputs meet organizational standards.

This feature addresses the challenge of low-density data from support tickets by generating synthetic Q&A from conversation streams. The synthetic data engine extracts meaningful questions and answers while discarding the original stream. This approach improves overall AI response quality without relying solely on sparse documentation.

The system includes proprietary deep AI models built on open-source large language models to perform specialized tasks. The synthetic knowledge extractor creates new knowledge from unstructured data. These models enhance AI capabilities and are exclusive to Twig customers, delivering advanced and domain-specific insights.

The platform uses retrieval augmented generation to reduce AI hallucinations and improve reliability. Answers are generated only from pre-indexed customer data, ensuring accuracy. Citations link back to sources, allowing teams to confidently share responses with internal or external stakeholders.