Though proliferation of novel democratic innovations continues apace, recent years have seen growing scepticism in the associated scholarship about their capacity to exercise genuine political influence. Critics highlight three particular concerns: that democratic innovations are seldom embedded in existing institutional architecture and political networks ; that democratic innovations are usually disconnected from, and potentially undermine, organic forms of bottom-up citizen participation; and that democratic innovations thus represent new tools by which elites manage and control an inauthentic form of public input. Critics typically present those engaged in implementing democratic innovation and public engagement as oblivious to these concerns, but in fact many practitioners are acutely aware of the obstacles and limitations that they face on the ground. This paper focuses on one such effort in practice to consciously confront and overcome the issues of embeddedness, connectedness and authenticity—the recent NHS Citizen process in England. We draw on focus group and interview data with the architects and organisers of the NHS Citizen innovation. We highlight the philosophy underpinning this innovation, the design strategies developed to put this philosophy into action, and the adaptation that occurred as the innovation was rolled out in practice. We argue that though NHS Citizen was not able to solve the perennial problems identified by sceptics in an especially hostile political context, it did pioneer a number of novel techniques and strategies that can contribute to learning in this field. We conclude by drawing out these lessons for the theory and practice of democratic innovation.