What Breaks When AI Goes to Production?
Part of Twin Cities Startup Week 2026, September 14–18.
Moving an AI product from prototype to production introduces problems that aren't always visible while you're building it. This panel will look at some of the most common ones: infrastructure that needs to support AI at scale, unexpected behavior and bias in AI outputs, cybersecurity and organizational risk, and the human challenges that come with introducing AI into the way people work.
The speakers will share what they're seeing in their own work, where companies tend to run into problems, and what founders and AI teams should be thinking about earlier in the development process. Attendees should leave with a better understanding of what to plan for before putting an AI product into the hands of customers or integrating it deeply into their organization.
Speakers & Perspectives:
Yossi Eliaz, AI Infrastructure Leader — Infrastructure: Yossi will speak to the technical infrastructure required to run AI agents at scale, including isolation, compute, concurrency, performance, and cost.
John Haggerty, AI / Cognitive Bias Expert — AI behavior: John's perspective focuses on what the AI actually produces, particularly cognitive bias, unexpected outputs, and issues that may not surface until a product is being used by real people.
Younouss Adjibi, Cybersecurity & Risk Leader — Risk and resilience: This perspective adds cybersecurity, governance, risk visibility, and operational resilience, including how companies can make more informed decisions about where and how they deploy AI.
Shannon Stauff, Organizational Change & AI Adoption Leader — People and adoption: Shannon will focus on what happens inside the organization when AI is introduced, including trust, resistance, capacity, agency, relationships, and whether the people expected to use the technology are actually positioned to succeed with it.
Getting here on Wednesday. Parking, Basecamp, lunch, and the rest of the day: tcstartupweek.com/wednesday