AI in practice, not theory, Emma Nguyen & Gary Bhanot | EvolveDigital Toronto 2026
Summary
Emma Nguyen and Gary Bhanot presented a practical framework for AI adoption at institutions, specifically drawing from their experience at the University of Toronto's decentralized marketing and communications hub. The session emphasized that AI adoption is primarily about change management rather than technology implementation. Emma outlined their four-phase approach: planning (securing leadership buy-in and establishing problem-first methodology), engagement (creating cross-functional working groups of 60+ team members to research AI solutions), enablement (conducting collaborative sessions and developing structured mini business cases for institutional approval), and scaling (building organizational confidence in trying new technologies).
The framework addressed common barriers including fear of job displacement, time constraints, and institutional approval processes. Gary detailed their technical implementation, focusing on their email marketing operations that send 1,600+ campaigns annually to 400,000+ recipients, generating $1.6 million in fundraising.
His team built custom GPT tools to solve scale challenges, creating a prompt-free environment with guardrails ensuring consistent outputs. Their custom GPT processes copy decks into HTML emails with proper metadata, tracking parameters, and brand elements across 30 university divisions. A second tool called 'Mailense' analyzes email effectiveness using historical performance data.
The speakers emphasized that success came from standardizing existing processes first, then applying AI to enhance efficiency rather than replace human creativity. Key outcomes included reduced manual work from 10 minutes to 3 minutes per email, elimination of human errors in repetitive tasks, and repositioning teams toward more strategic value-added work. The approach enabled teams to overcome AI hesitation by focusing on solving actual workplace problems rather than adopting technology for its own sake.
Key Takeaways
- •Start AI adoption with leadership buy-in and documented plans rather than jumping into tool experimentation
- •Use a problem-first approach - identify workplace problems before seeking AI solutions, not the other way around
- •Create cross-functional working groups to research AI applications specific to different roles and expertise areas
- •Develop structured mini business cases to navigate institutional approval processes for new AI tools
- •Build prompt-free AI environments with guardrails to ensure consistent outputs regardless of user skill level
- •Standardize existing workflows and collect metadata before implementing AI to maximize effectiveness
- •Focus on augmenting human capabilities rather than replacing jobs to reduce adoption resistance
- •Allow teams to identify their own problems rather than having leadership dictate what needs solving
- •Implement AI to handle repetitive tasks so teams can focus on strategic, value-added work
- •Create opportunities for cross-team learning where different roles can share AI applications and inspire new ideas
- •Document the entire adoption process to build institutional confidence and replicability
- •Prepare for lengthy institutional approval processes by having multiple AI initiatives in parallel pipelines
Transcript
Hi everyone. Uh well, welcome everyone to the first talk here. Thank you so much. Uh my name is Anna. I'm the senior account manager with evolving web and I'm here I have the pleasure of introducing the next talk and the next speakers Emma and Gary and their talk on AI and practice not theory. So I have like a little summary. So what does it actually take to get to using AI? not just experimenting with it but truly integrating it into everyday work. And that's the question at the heart of today'...
