The AI Panel's Rule: Start With the Problem, Not the Tool

On stage: Julie Souza, tech and innovation advisor and former global head of sports, AWS · Samantha Stark, founder and chief strategist, Phyusion · Moderator: Becki Wronko Civello, COO, Next League

Problem first

Souza's rule is that AI should be a tool for hitting business goals, not a box to check. Ask what problem you're solving, what opportunity you're seizing, what outcome you expect and whether you can measure it. She remembered when every company suddenly needed a metaverse strategy.

Her favorite example is the NFL, which she said used 10,000 seasons' worth of synthetic data to rethink the kickoff, a play that was rarely returned and carried far higher concussion and injury rates than others. The result was the league's dynamic kickoff.

She also described a sports team that set up multi-agent systems to turn its data into content, producing far more of it at a fraction of the cost. "And this isn't displacing anybody," she said. The team could never have staffed that much content. What it unlocks is personalization.

Rebuild the workflow

Stark said she founded Phyusion because she lived through the arrival of the internet in her early 20s without being in a position to make the most of it, and promised herself she wouldn't miss the next big shift.

The common mistake, she said, is layering AI on top of how you already work. Her firm starts by diagnosing every workflow and where it gets stuck, as it did with a team at New York Life, its first client. It also builds synthetic audiences: AI replicas of consumers, a boss or investors that you can test messages against before the real thing.

Teams that learn to use it well get time back, she said, and many can take on projects they couldn't before. The skills are new, too. "That's a new skill, context engineering," she said, meaning knowing what an AI needs from you to give you something real. In her last job she oversaw 400 people. "Now I am like 10 humans managing 40 agents."

One panelist said she taught herself the creative tools two years ago by making her holiday card from scratch with generative AI, from the animation to the voiceover to the music.

The other side of the ledger

Wronko Civello pushed the panel on concerns. They named data centers first: energy and water use, how many jobs they really create and what they do to local communities. Like social media before it, they said, the companies profiting need to pay their share. They also raised reports of AI agents slipping their guardrails in testing, and the risk of letting AI do your thinking. Writing something down, one said, is how you actually think about it.

Prove the value

On change management, Souza told a story from early in her career, when she was implementing enterprise software. One employee refused to use it. Souza timed her doing a calculation by hand, about a minute and a half, then had her push the button. The employee started to cry, Souza said, and became the system's biggest evangelist.

"Tech for tech's sake is terribly boring," she said. Prove the value first.

Asked whether AI has made her a better leader, Stark said yes, "because it's removed all of the admin work." The panel's advice for getting started: pick one recurring task with a set structure, like a weekly report, and go deep on it.

The Coach's Board take

Youth sports runs on volunteer hours, which makes this advice practical. Pick one recurring chore, like the weekly schedule email, registration reminders or the game-day volunteer list, and automate that first. And keep players' and families' data away from any tool you don't fully control.