Eighty-five percent of companies are running AI proofs of concept. Only 5% make it to production.

Aaron Rallo, co-founder and CEO of Trovia and former leader of agentic AI at AWS, breaks down exactly why and what the successful 5% do differently.

The problem isn't the technology. It's content quality, governance, and designing for scale before writing a single line of code.

Aaron shares the "consequence and confidence matrix" he developed at AWS to determine when AI should act autonomously and when a human must stay in the loop, why the "accept all" pattern is the most dangerous habit in AI deployment, and how to anchor every AI initiative to a concrete million-dollar business problem before starting a POC.

Most organizations plan for prevention and call it resilience. Mark Orsi says that's the mistake that gets exposed the moment something actually breaks.

Mark Orsi, CEO of the Global Resilience Federation, spent years as a data protection leader inside some of the largest banks in the world before concluding that even billion-dollar security programs were never going to be enough on their own. In this episode of Tools for Tech Leaders, Mark breaks down the Operational Resilience Framework his organization built with hundreds of companies, walks through the real story of how the industry coordinated its response to the CrowdStrike outage, and shares a finding from a 700-bank payments exercise that revealed a massive, previously invisible gap between what banks assumed their vendors could handle and what those vendors could actually deliver.

The conversation also covers AI as an emerging category of critical infrastructure, the identity risks created by deepfakes and autonomous agents, and the specific communication habits that separate leaders who keep customer trust through a crisis from those who lose it.

This episode provides a concrete framework for executives who need to plan not just for prevention, but for what happens after prevention fails.

Deb Liu spent years pitching Facebook Marketplace before anyone would greenlight it. When it finally launched, it took multiple failed versions before reaching a billion users. That arc — patient conviction, rapid testing, willingness to be wrong — is the throughline of everything she shares in this episode.

From her time building payments at PayPal and ads at Facebook to leading Ancestry as CEO, Deb has operated at a level of scale most product leaders never reach. She breaks down what actually makes marketplaces work, why the transition from IC to manager breaks talented people, and what the CEO job demands that no one prepares you for. She also shares a sharp framework for how to think about AI adoption right now: we are, in her words, in the 1998 of the internet.

A must-watch for product leaders, founders, and executives who want to build things that last — and lead teams that can actually get them there.