Canvs Upgrades AI Story Assist with Focus on Trust

Text analysis platform Canvs AI has announced a new release of its AI Story Assist software, incorporating a new executive summary feature with interactive citations and quantification of the themes, topics and emotions identified in open-ended text data.

New York-based Canvs combines AI Story Assist - launched a year ago - with patented natural language algorithms, to deliver insights from unstructured feedback - including quantification of themes, topics and emotions - and convert open-ended text from surveys, social media and online reviews into a natural language summary. Researchers can then use AI Story Assist to ask natural-language questions and drill into quantifiable analysis to further explore and validate insights.

The firm acknowledges that there is widespread mistrust of AI outputs. 'Insights leaders need to understand the source of any AI-generated insight to trust their reputation on it', comments Kristi Zuhlke, Canvs' Principle of Generative AI for Insights. Addressing this, the firm says it has 'prioritized trust, transparency, and privacy within AI capabilities' since launch in 2018, and underlines this commitment with features of the new version including Interactive citations, allowing researchers to verify the authenticity of AI-generated summaries and takeaways; a closed-domain approach, ensuring privacy and that all outputs originate solely from the users' data source; and connected analysis, linking efficient gen AI outputs with other, 'robust' insights discovery tools.

According to founder and CEO Jared Feldman (pictured): 'Without trust, insights generated from AI are worthless. That's why Canvs AI has prioritized transparency in our AI-powered insights platform from day one. This release marks an industry milestone in high-trust AI for insights, building citations directly into the AI Story Assist executive summary generated on every project, along with the ability to ask custom questions and drill into quantified analysis of open-ended feedback'.

The company is online at www.canvs.ai .

All articles 2006-23 written and edited by Mel Crowther and/or Nick Thomas, 2024- by Nick Thomas, unless otherwise stated.