In partnership with Dataiku and the EDM Association, Bronson was pleased to host “AI for Executives Breakfast: The Future of Work with AI” at the Royal Ottawa Golf Club in Gatineau on Tuesday, April 21, 2026. Convened for senior public sector leaders, the morning offered a focused look at how AI is reshaping executive work, and at which capabilities remain durably human as the technology matures. Attendees gained both a strategic and practical view of the conditions under which AI initiatives succeed in enterprise settings.

Martin McGarry, President and Chief Data Scientist of Bronson, opened the session by putting a single executive brief in front of five contemporary models: Cohere Command R+, Microsoft 365 Copilot, Gemini 3.1, Claude Opus 4.6, and GPT 5.4. Identical inputs produced materially different results, and the response that ultimately met the standard required roughly fifteen minutes of expert iteration. Martin used the finding to frame the central thesis of the morning: the model is no longer the constraint; the expertise applied around the model is.

Glendon Hass, who leads Bronson’s AI and Automation practice, picked up the question raised by Martin’s demonstration: if the model is no longer the constraint, what is? His answer identified three durably human capabilities: judgement, expertise, and taste. He traced how the human-AI workflow shifts from yesterday’s fully human practice, through today’s hybrid model, to a future state in which human judgement bookends a largely machine-run middle. Glendon walked the audience through a representative Treasury Board submission, demonstrating how the same model and the same scenario can yield either a generic baseline or a submission defensible at the Secretariat which depends entirely on the scaffolding around the work.

Doug Harrison, Country Manager at Dataiku, spoke regarding “What Actually Drives AI Success: Beyond the Hype.” Drawing on decades of experience helping enterprises scale data and analytics, Doug argued that most organizations are not failing at AI because of the models, but because of everything around them. His remarks focused on three pillars critical to enterprise AI success: people, orchestration, and governance, and on the operational shift required to move from fragmented pilots to repeatable, enterprise-grade AI delivery.

The morning closed with a panel discussion moderated by Glendon Hass and featuring Paul Childerhose of the EDM Association, Lyudmila Hamka, Director of Operations Intelligence at Canada Post Corporation, and Hosai Zurmati-Halim, Senior Policy Advisor at the Treasury Board Secretariat. The conversation focused on the organizational, cultural, and policy realities of AI adoption.

Thank you to Doug Harrison and the team at Dataiku, to Paul Childerhose and the EDM Association for their partnership, to Lyudmila Hamka and Hosai Zurmati-Halim for sharing their perspectives on the panel, and to the public sector executives who joined us for a generative morning of discussion.