Using a real-life and imagined case studies, we demonstrate how a casual, informal political conversation on social media among ordinary citizens could be transformed into a policy discourse. It is done by deconstructing the logic of discursive interactivity of online discussions. In doing so, we apply Jurgen Habermas' validity claims to normative rightness to reveal citizens' attitudinal positions 'For' and 'Against' certain social effects of food destruction policy of the Russian government. We measure citizens' attitudes in the form of the discursively constructed solidarities behind each position and show the interactive process of their formation. We also build a range of interactivity models that could exploit the potential of artificial neural networks for creating new tools of discourse analytics that can capture citizens' policy inputs in an easily understood format. The goal of such tools is seen in helping reduce deliberative disagreements by encouraging acceptance of other points of view-the core principle and ideal of deliberative democracy.