As AI hype wanes, substance beats showmanship as experts demand measurable impact

As the market matures, AI hype is giving way to hard questions. Integration, ROI, and real-world value are in the spotlight as enterprises grow wary of promises without proof. The conversation is no longer about what AI could do, but what it actually delivers.

Mike Haylon, General Manager of Asana's AI Studio, builds with an enterprise sales mindset and zero illusions. AI agents may be the future, but according to Haylon, most companies in 2025 aren’t ready. Engineering hurdles and overhyped expectations still stand in the way of real deployment.

CIOs want receipts: Standing out is about substance, not sizzle. After speaking with CIOs across Europe and North America, Haylon has a clear directive: "Don’t bother with a differentiated message. Show me a use case. Show me a before and after of the impact that you can drive."

That demand for substance also tempers the hype around AI agents. "There’s a lot of talk about agents," Haylon says. "While we believe too that agents will be indispensable, right now there’s a challenge in deploying them really effectively."

Data indexing dangers: Haylon flags a common trap: companies assume that dumping all their data into one place and layering on search will make agents work. "Just indexing all your data into a single place and deploying agents doesn't mean they'll be effective," he says. "There’s not much strategy in that. It just frustrates and confuses the agents and the people behind it."

Structure beats spectacle: Asana skips the flashy demos and starts with the basics. "We’re focused on a clear business case that’s rooted in an existing manual workflow" Haylon says. Their workflow builders act like GPS, giving agents "point by point turns" to follow. "We’re giving you directions to where you’re going," Haylon explains. It’s a methodical, structured approach—and most importantly, it works. AI performs best when anchored to real processes and clean, purposeful data.

Power to the people: For Haylon, real AI value comes when non-technical users can actually use it. "You can give them clear context, you can give them clear direction," he says. Asana’s workflow builder is designed to be so intuitive that "we're able to empower non-technical people to learn reasonably quickly how to take advantage of AI." What makes it work? "Context, structure, and guidance—exactly what's required to get the most out of AI as it exists today."

Show, don’t tell: Asana’s strategy leans on education and real-world proof. It’s about "education, enablement, and giving people hope and belief in how their use case can be better supported by AI," says Haylon. While LLMs may deliver fast answers, true value takes more. "Our job is to help bring that to life by telling stories around the use cases we are deploying successfully." The way to differentiation, he says, is "putting it in customer's terms and then really quantifying it when you go to the CIO."

Pragmatism prevails: AI agents may dominate the conversation, but Haylon isn’t sold on the hype cycle. "It's clear that not everybody knows exactly what it means. There are more practical ways to get value out of AI than broad deployment," he says. What matters most, he adds, is execution: "Some structure and some guidance along the way goes really far," says Haylon.