WEB SUMMIT VANCOUVER 2025 || From Hype to Hook: Why AI Needs Purpose, Not Promotion

With every other tech startup and business now “AI‑powered,” the real question becomes: is it solving something, or just checking a box?

Opening with Scale—and a Caution

Web Summit Vancouver welcomed 15,727 attendees from 117 countries, showcasing a first-year record of 1,108 startups—a tidal wave of innovation and energy that quickly becomes overwhelming. The event was structured as a choose-your-own-adventure format, with tailored summit tracks (New Media, New Venture, Commerce, Corporate Innovation, Marketing, Growth, Creative, AI, and People) to choose from. Going through the schedule takes quite a bit of time, and building out your schedule to ensure you cover topics of interest will take around two hours. Of course, you should also build in time to attend some of the meetups and network amid all the learning and nuggets of wisdom you’re trying to absorb.

I selected a combination of events from the Marketing and AI Summits to ensure I could build upon my strengths and widen my understanding of what’s on the horizon with the rapid evolution of AI.

The Momentum Trap: AI for the Sake of AI

It wasn’t surprising that every second booth claimed to be “AI-powered,” and I was glad that the sessions I attended questioned the depth behind the label. Co-founder & Managing Partner at FPV Ventures, Wesley Chan, during “Smart Money in 2025,” reminded us that real impact comes from solid business fundamentals. Canva, for instance, succeeded not because of its AI but because it democratized design. Its AI features were helpful, not essential.

The Real Value Test: Purpose‑Driven AI Adoption

1. Smarter, Not Harder

In his fireside chat, Ping Wu from Cresta (moderated by Chris Marr of PwC), titled “Smarter, Not Harder: AI’s Role in Data‑Driven Optimization”, emphasized that AI should optimize, not overcomplicate. Focus on meaningful results, not flashy add-ons.

2. Data, APIs & Change Management

Wu urged businesses to invest in clean data, robust APIs, and change management for both employees and customers. The goal isn’t an AI gimmick, but a foundation that allows AI to automate intelligently, from infrastructure up to real-world applications.

3. Autonomy with Intention

Darin Patterson, VP of Market Strategy at Make, hosted “How to Build and Manage AI Agents Without Code.” He urged a shift: agentic AI, or systems that can autonomously act within dynamic environments. He warned against building agents purely for show. Instead, they should be rooted in real-world workflows, not flashy demos.

Framework: From Hype to Hook

Here’s a 5-step framework I suggest to avoid adding AI to your business just because it’s the latest shiny object.

1. Define the problem first. Is there a clear, valuable issue to solve? Many businesses are adding AI like it’s a search keyword they’re inserting into a blog article on their website.

2. Check business foundations. Would the solution stand without AI? In other words, will your service or product function with the same level of efficiency and quality with or without the addition of AI? This is likely a red flag, indicating that you may not need AI for your business.

3. Build the stack. Ensure clean data, accessible APIs, and manage internal and external expectations. This is what I learned from Ping Wu. You’ve got to ensure it’s not garbage in, garbage out with your data and processes. It’s also crucial to have conversations with all stakeholders to ensure everyone is aligned with the expected results (and when the results aren’t as expected, management should step up to address it and have a backup plan).

4. Pilot with no-code agents. Let experimentation lead to refinement and adoption. Here’s what I learned from Darin Patterson’s session to assess whether you should delegate a task to an Agent. 1) Does the task require complex decision making? 2) Are there difficult-to-maintain rules? 3) Is there a heavy reliance on unstructured data? If you say yes to any of these items, you could consider building an agent to assist with the overall workflow. The example Darin used was for an e-commerce website with a lot of returns. The agent was provided with all the rules and policies to review and process returns, and flag outlier cases for further human review.

5. Measure and validate. Are you improving metrics that matter? Integrating AI into your business solutions isn’t a light lift for your team. You must assess whether the input is worthwhile for different business horizons. I would argue that businesses should only consider investing fully in AI when they fully understand how the technology can help differentiate or improve the solutions offered in the medium to long term.

AI framed as a buzzword drains attention and resources. True innovation emerges when businesses ask: What problem are we solving? That clarity distinguishes transformational uses of AI from symbolic ones—and avoids chasing the next shiny trend.