How do private interests try to shape public interest competition regulations? Focusing on debates about the design of wholesale Internet access obligations, the authors employ Natural Language Processing (NLP) tools to evaluate a multi-stakeholder policymaking process in Canada. Using NLP, they analyze 40 formal interventions in the CRTC's 2013–551 review of its wholesale broadband policy. They classify major interest groups, map key concepts, and quantify asymmetries in stakeholders’ influence. They conclude that by reducing the costs of regulatory participation, deploying NLP technologies can help offset the advantages large incumbent organizations already have in shaping law and policy.