Splice CEO warns AI-generated emails erode workplace conversation quality

Kakul Srivastava, CEO of the sample platform Splice, argues that AI-written emails and documents degrade workplace conversations by obscuring human intent. In an interview with The Verge, she states that once the origin of a document is unclear, the subsequent discussion becomes less productive and harder to contextualize. 'Almost everything good that’s happened here started' with real conversations and well-written documents where authorship is evident, she said.

Srivastava has emphasized keeping humans central to AI tools at Splice, including its generative AI offerings. She previously criticized 'push-button, get-song' AI approaches associated with companies like Suno, advocating instead for tools that preserve creative agency. 'I think the same thing is true of music,' she added, suggesting that opaque AI outputs risk undermining the collaborative process in creative fields.

The executive also described Splice’s remote-first culture, where collaboration relies on asynchronous, written documents and structured debates rather than hallway conversations. 'We don’t have a hallway. What we have instead is somebody sitting down and actually thinking something through, and then a bunch of people arguing with it,' she said. This structure makes clear authorship and human reasoning especially critical, she argued.

Srivastava’s comments come as Splice has expanded its platform and explored generative AI, including the acquisition of Spitfire Audio. The company positions its AI tools as adjuncts to human creativity rather than replacements, a stance that contrasts with platforms offering fully automated outputs.