Opening its doors for business this month, US-based Kumkuat AI is designed to help companies working in corporate communications, investor relations and public affairs, enabling them to “evaluate, test and optimise” ahead of going public with announcements – ascertaining how audiences might react.

Kumkuat AI said its platform will address three core needs: what the most important stakeholders think about a company and its executives, why they have those particular opinions and how to best respond in terms of actions and messaging.

It is designed to cut through the need to manually analyse millions of datapoints to glean salient insights. Instead, it constructs data-informed synthetic versions of audiences based on profiles of people, organisations and groups that influence corporate reputation.

Kumkuat was founded by Dan Gaynor, an executive fellow at Harvard Business School; former Publicis Groupe chief strategist Shann Biglione; and former KKR data infrastructure expert Dylan Roy.

Gaynor said: "The savviest communicators prioritise one thing above all else: how key audiences feel about what you do and say.

"But until now, it’s never been possible to preemptively test every move or message. At a time when valuations and sales can swing on sentiment alone, this is about more than strengthening reputation; it’s about accelerating growth. We're enabling communicators to validate their instincts with evidence – in seconds, not days – and in doing so, elevate their influence."

Kumkuat has been built on the idea that synthetic audiences can be just as reliable and hugely more detailed than real ones. Customers can use it to hold live conversations and prepare for tough questions from the likes of media; generate scripts, announcements and press releases; and share data across teams in various formats, allowing feedback.