Takeaways from World Summit AI Montreal 2025

The prevailing wisdom among the presenters and panelists at the recent World Summit AI conference in Montreal can be reduced to these two statements:
‍

• All organizations must adopt AI. If they don't, competitors will either outprice them thanks to productivity gains or outcompete them on customer/member experience.
• Tools don't replace people. People who use tools replace people who don't use tools.

‍

In short, both organizations and their people are at risk of being outcompeted. In some industries this will come rapidly, and in others it will be more of a long-term adaptation.

But there are many challenges to be overcome by all organizations that want to emerge from this wave of change at the top of the food chain. Here are some of our key takeaways from the event:

‍

Adoption Must Be Prioritized

Great tech fails if people won’t use it.

The word adoption probably came up more than any other over the course of the event. Of course this is nothing new - at Differly we help organizations across industries with technology selection and implementation, and adoption is baked into our process from start to finish for good reason.

But AI does have some novel features that make it particularly tricky: prompting, open-ended outputs, different models for different use cases, not confusing it when you provide reference documents or include past chat histories, and of course there are concerns about information and job security. Here's what we heard about adoption:

‍

Employees abandon AI tools when their first try isn’t obviously useful.

Employees are busy, with lengthy to-do lists and meetings to attend. For most people, if an AI tool doesn't provide an immediately useful solution to a problem, they'll continue doing things the way they always have.

The most common first step that organizations take when rolling out AI is to give a subset of employees Copilot licenses. This often hinders AI adoption. Why? Because Copilot uses less powerful models that tend to produce generic outputs that aren't all that useful.

Absolutely give your most innovative problem solvers access to licenses (with an appropriate usage policy) so that they can discover new solutions to high priority problems. But for most employees, their AI journey needs to be thought through more carefully:

• Focus on high priority workflows to important problems and make sure the AI tools reliably deliver useful results before rolling them out more broadly.
• Think about the interface. If you're asking your employees to use many new interfaces across many new apps, they are more likely to feel frustration and confusion.
• Provide hands-on training. People need to use AI to get good at using AI.

‍

Taking Informed and Managed Risks

AI technologies come with inherent information security risks. Sensitive or private information may be exposed to the technology for use in its training, meaning your data becomes part of the system's 'knowledge' and could appear as outputs of the system to users outside of your organization.  Incorrect outputs may be relied upon as being true, with potentially severe consequences.

Leadership teams need to understand the risks and make informed decisions about what level of risk they are comfortable with. Without a clear AI usage policy, organizations face two likely outcomes:

• A wild, growing tangle of unvetted AI-based tools integrated with and exposed to your data
• A full stop, with IT shutting down all AI usage with overly restrictive policies.

‍

Top Use Cases for SMEs

While we were dazzled with presentations detailing how enterprises like Spotify and Wayfair are using AI, there is a lot of low hanging fruit for SMEs that won't require major IT investments and platoons of data scientists:

• Reviewing documents like policies, proposals and contracts for errors, compliance and readability.
• Automating routine administrative tasks like approvals and project coordination
• Converting data from one format to another, such as from unstructured pdfs to structured data tables.
• Preparing for and summarizing meetings

‍

And of course, these lists of practical use cases tend to forget about what the research is showing is the most valuable benefit of AI for knowledge workers today: having a thinking partner to plan, ideate and solve problems with.

‍

How Differly Can Help

Get in touch with us today to learn more about our AI Executive Briefing. We will get you started on your journey of adopting AI-fueled advancements towards your strategic goals. We take the technical complexity out of the conversation, focusing on priority business needs and the people who ultimately make change happen.