It began with efficiency: how can generative AI help work get done faster, with less effort? Then came scale, as professionals unlocked the ability to produce more, at speed.
Today, as generative AI becomes increasingly embedded in everyday work, the bar is again rising – from productivity to value, and from output to sharper thinking and greater creative potential. So, what does it take to get there?
That question was at the centre of Impact Live – Beyond the prompt: Generative AI at work, where Ivey’s Dean Julian Birkinshaw, MBA ’91, PhD ’95, brought together three voices working at the forefront of AI in practice: Fab Dolan, HBA ’08, Founder and CEO of 99 Ravens AI; Alyza Keshavjee, HBA ’06, Head of Consumer Packaged Goods at Google; and Mazi Raz, MBA ’05, PhD ’14, Ivey Assistant Professor of Strategy.
What emerged from the discussion wasn’t a list of better prompts or tools, but a reframing: generative AI cannot create value on its own. It amplifies the thinking and judgment that individuals bring to it. The question, then, is not how to use AI better, but how to think with it, and how to do so in a way that creates real advantage.
What follows are five insights from the panel to help you do exactly that.
1. Move beyond the prompt
For many professionals, using generative AI effectively starts with prompt engineering – the practice of crafting questions to guide an AI system, like ChatGPT, toward more accurate or useful outputs.
But Dolan cautioned that this approach introduces organizational inconsistency at scale. “If everybody is prompting differently,” he noted, “you’re going to get different outputs.”
Instead, he encouraged a shift toward context engineering: shaping the information, background, and framing around a task so AI can produce more relevant, consistent, and high-value results across teams.
Put simply, prompt engineering is about asking better questions. Context engineering is about creating the conditions for better answers.
2. Get the timing right
One of the most important shifts in getting real value from AI isn’t how you use it, but when you use it.
For Raz, that begins even before the prompt is written, with users pausing to examine and question their own thinking, intentions, and judgment. As he put it, “it’s not just about seeking an output, but knowing what a good output should look like.”
Dolan sees timing as equally important. Too often, AI is brought in too late and used only to generate final outputs. But when it is introduced further upstream – while ideas are still forming and decisions still open – it can challenge logic, surface blind spots, and strengthen thinking before it scales.
3. Think with AI, not just through it
It’s tempting to treat AI like a shortcut: ask a question, get a polished response, move on. But that approach risks shallow thinking, embedded bias, and misplaced confidence.
Raz offered a cautionary metaphor: “to me it’s almost like a vending machine.” The output may look complete, but users often don’t understand what went into it, or what’s missing.
The professionals who get the most from AI don’t just generate outputs – they interrogate them. They ask what assumptions are embedded, what perspectives are missing, and whether the answer actually holds up.
Used this way, AI becomes not just a generator of answers, but a partner in refining them.
4. Strengthen your human advantage
Paradoxically, some of the most valuable uses of AI are deeply human.
Keshavjee described using AI to rehearse conversations, pressure-test ideas, and prepare for high-stakes interactions. For her, the goal isn’t to generate a ready-made response, but to perform better when it matters most.
This philosophy is woven into her team’s day-to-day work: “The default expectation with my team is everything should start with AI.”
Not because AI should do the work, but because it helps professionals show up more prepared, more thoughtful, and more effective in moments that require empathy, persuasion, and alignment.
5. Experiment deliberately
If there was one practical takeaway shared across the panel, it was this: experiment deliberately.
“Test one thing a day differently,” Keshavjee advised. Done consistently, it shows where AI can make a meaningful difference – and where it can’t.
And, that last point is critical.
As Raz noted, AI is always available, but it shouldn’t always be used. Knowing when to lean in and when to hold back is part of the new discipline of working with AI – and it’s built through deliberate, strategic experimentation.
Where real value comes from
What emerged from the discussion is a clear pattern: the greatest gains from generative AI will not come from doing more, faster, but from working more intentionally.
For professionals, that means using AI as a partner earlier in the process, applying enhanced critical thinking, and remaining actively engaged in – and skeptical of – the work it produces.
Because, ultimately, who gets the most value from generative AI won’t be those who use it most, but those who bring the most to it.
Watch the full Impact Live webinar, Beyond the prompt: Generative AI at work, on Ivey Impact or Ivey’s YouTube channel.
