The Real Reason AI Projects Fail, According to Prezi’s CEO

For years, leaders have been told that artificial intelligence is the competitive edge. According to Prezi CEO Jim Szafranski, that thinking is backward. “The technology is not the hard part,” Szafranski said. “Finding the right problem, that’s the hard part.” Most companies are getting that wrong. The myth of starting with technology Szafranski explained leaders often begin their AI journey by asking, “Where can we use AI?” instead of “What are we actually trying to fix?” That misstep is costing companies billions. According to Gartner, as many as 50% of AI projects fail to deliver meaningful results, largely due to poor alignment with business goals. Szafranski saw this play out in a steel mill project. “We thought we were solving for scheduling, but that wasn’t the real issue,” he said. After deeper analysis, the team discovered the real problem was optimizing how steel reached customers, not replacing a human scheduler. Once reframed, the AI delivered actual business impact. “The first problem you see is almost never the right one,” he added. Finding the “perfect problem” Szafranski described what he called the “perfect problem,” a challenge that is both meaningful and solvable. “You’re looking for something where the impact is obvious, and the path is achievable,” he said. “That’s where AI works.” AI pilots fail to produce measurable business impact, not because of weak models, but because companies pursue the wrong use cases. The takeaway: AI success is less about sophistication and more about precision. Why “time to outcome” beats “time to value” One of Szafranski’s biggest shifts in thinking is moving beyond “time to value.” “Time to value is incomplete,” he explained. “What matters is time to outcome, did the user actually achieve what they needed?” That insight reshaped Prezi’s AI strategy. Initially, the company focused on automating presentation features, making slides faster and easier to build. However, that wasn’t the real job customers needed done. “They’re not trying to make slides,” Szafranski shared. “They’re trying to persuade somebody.” That realization changed everything. What Prezi is doing differently Today, Prezi is using AI to help users communicate and persuade more effectively, not just design better presentations. “We shifted from helping people build presentations to helping them win moments,” Szafranski explained. The platform now focuses on: Simplifying visual storytelling for non-designers Helping users communicate ideas quickly under pressure Enabling more engaging, outcome-driven presentations This shift has unlocked growth, particularly in global markets. Szafranski noted that accessibility has become a major driver. “When you remove the barrier of design skill, you open the door to entirely new audiences,” he said. That strategy is working. Prezi continues to expand internationally, especially in regions where traditional presentation tools were harder to adopt due to language or educational barriers. Accessibility is a growth strategy, not a feature Prezi’s approach highlights a broader truth: accessibility is inclusion and expansion. According to MIT research the vast majority of AI investments fail to generate financial returns when they are disconnected from real user needs. Prezi is doing the opposite — building for real-world communication challenges at scale. The real takeaway for leaders AI isn’t magic. It’s a multiplier. As Szafranski made clear, “If you pick the wrong problem, AI just helps you get there faster.” The companies winning with AI aren’t the ones with the best models. They’re the ones asking better questions. Because in the end, the difference between failure and transformation comes down to one decision: Are you solving the problem you see or the one that matters? BY NETTA JENKINS, FOUNDER, HIC; WORKPLACE CONSULTING FIRM | AUTHOR OF SUPERCHARGED TEAMS