Only 13% Can Measure AI ROI. Now What?

Only 13% Can Measure AI ROI. Now What?

When only 13% of companies feel confident measuring AI ROI, you know two things: you are in a hype cycle for a new technology, and operational issues prevent progress. AM Best’s report, “Artificial Intelligence Appears to be Ready, But Most Insurers Are Not,” makes that reality hard to ignore. Sure enough, 45% of insurers cite data readiness as a top challenge. For those of us who have spent years saying “it’s all about the data,” it lands with a thud. The “told you so” moment feels tempting. It also does not help. Data readiness is a problem across every industry, and insurance carries extra complexity because of legacy, regulatory oversight, and the need to defend decisions that affect consumers. Why there’s a gap between pilots and board-ready outcomes should be interrogated, and this episode delivers insight. You’ll gain a pragmatic view at how both organizational and industry dynamics significantly impact success or failure. Two industry veterans offer a front-line view: Jeff Rieder, Partner and Head of Benchmarking at Aon, and Stefan Holzberger, Executive Vice President and Chief Operating Officer at AM Best. Jeff brings a benchmarking lens on how insurers evolve through tech cycles, where job families shift, and why executive alignment and measurement determine whether adoption sticks. Stefan brings AM Best’s lens on innovation, stability, and risk. He shares where insurers deploy AI first, why claims and back-office workflows move faster, and why underwriting adoption demands governance discipline and regulatory awareness. Leadership, culture, and talent development emerge as the common thread. AI does not move through an organization on its own. Companies need leaders who set direction, teams who build foundations, and talent strategies that expand skills instead of amplifying anxiety. It is no surprise that ROI confidence remains low when organizations still struggle to connect data readiness, governance, and adoption behavior to measurable outcomes. Three takeaways from the episode: * Treat data readiness as an operating priority, not a side project. * Define board-ready success measures early, then manage to them with leadership alignment. * Build governance and talent development in parallel so adoption scales without breaking trust. Thank you for listening. - Kirstin This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit leadthemachine.substack.com [https://leadthemachine.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Systems Thinking for AI: How to Get Results Faster

Hi everyone, The rubber meets the road when we need to prove ROI with any new technology. It’s inevitable and it’s hard. Rob Cressy is an AI enablement coach who works with leaders and teams that want measurable performance from AI. I like that he brings a human-first approach, and anchors AI adoption in identity, vision, and values. Then turns that into execution through systems. AI results come from systems, not dabbling Rob and I talk about a pattern I see across industries. Teams treat AI like a search box. They ask a question, accept the first output, and stop. When the output disappoints, they blame the tool, the data, or the moment. Rob calls this a foundation problem. He says: “the first prompt is the start, not the end.” He also gives a clear reason executives should care. AI adoption creates performance spread inside the same team. Rob shares an example where one person doubles output using AI. That person creates a gap that compounds when peers stay on old workflows. Leaders feel the impact in slower delivery, lower win rates, and more friction across the organization. We also talk about what leaders can do right now. Rob recommends a simple discipline. Teams should list daily friction, choose one workflow to improve, then ship a small win. He pushes leaders to build a roadmap from the work people already dislike, and he reinforces a systems lens. Systems scale, and clarity scales. Rob offers one prompt I expect to reuse: “What is hidden and non-obvious?” He uses AI to surface blind spots, simplify systems, and reduce unnecessary complexity. He treats AI as a tool that supports structured execution, not a replacement for leadership. If you lead a team through AI adoption, this episode will help you build a foundation, improve output, and keep the work human. — Kirstin This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit leadthemachine.substack.com [https://leadthemachine.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Talent Strategy in the Age of AI

Hi everyone, This week on Lead The Machine, I recorded a special dual interview on site at the Emerging Leaders Conference in Nashville with Wendy Davis Johnson and Marguerite Tortorello. We used the moment to look backward and forward. We traced the early days of the Insurance Careers Movement (ICM), and we connected that work to the talent and skills demands AI creates right now. I’ve shared before that I helped co-found ICM, now in its 11th year. It’s one of my proudest accomplishments. Wendy, author and corporate strategist, took us back to the beginning. She described how she and Brian Duperreault [https://leadthemachine.substack.com/p/ai-is-inevitable-relevance-is-a-choice] looked for a topic that mattered, researched diversity and leadership visibility, and then found the deeper signal. The industry faced a looming talent gap as experienced professionals retired and fewer young people entered the field. Wendy described how that research led to a simple conclusion. The industry needed a louder voice, and leaders needed to treat talent as a strategic priority. Marguerite, Executive Director of ICM, described what happened next. ICM grew from humble beginnings into a global collaboration. Today, 22 countries participate in Insurance Careers Month, over 1,000 organizations engage in the initiative, and the Emerging Leaders program has produced more than 1,000 alumni across functions, roles, and career stages. Companies now treat February as a kickoff for year-long talent planning, not a one-off recruiting campaign. We also talked about AI. Headlines focus on layoffs and replacement. I see a different path for winning. Organizations win when they invest in people, build AI fluency, and connect that fluency to real customer and operational outcomes. Marguerite described how companies already run intentional upskilling efforts, and she highlighted the role of legal, compliance, government affairs, and regulators in responsible adoption. Wendy reinforced the core reality. AI accelerates research and analysis, and leaders still differentiate through human judgment and relationships. If you care about talent strategy, leadership development, and the human side of AI at work, this episode will give you history, context, and a grounded view of what winning looks like. — Kirstin This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit leadthemachine.substack.com [https://leadthemachine.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

AI Tools That Reduce Burden and Restore Dignity

Hi everyone, This week on Lead The Machine, I spoke with Dakota Koontz, Executive Director at Housing Heroes Hub, about what it takes to lead inside broken systems, and how AI can reduce the load for leaders who carry too much. Now… this may seem counter to my consistent advice to fix broken processes first, rather than slap AI onto inefficient workflows. This is different. Dakota works with housing leaders who face rules layered on rules, outdated technology, and constant pressure to stay compliant while serving vulnerable communities. He also names the part leaders rarely say out loud. The job includes heavy emotional labor. Leaders manage tenants, staff, funding risk, and public scrutiny, often at the same time. These leaders aren’t in a position to fix government systems and regulations. They must adapt, ensure compliance and advocate for their constituents - while being overworked. We focus on that business problem and talk about the technology that can help. Dakota shares concrete ways teams use AI: * Convert long intake packets into digital forms. * Reduce back-and-forth with cleaner inputs. * Prepare for board, funder, and stakeholder conversations with mock dialogues. * Improve prompting with a standard operating procedure (SOP) mindset so output stays consistent. He also makes AI adoption accessible. He does not come from a traditional tech path. He invested the time, built skill through repetition, and taught thousands of people how to do the same. This episode focuses on clarity, dignity, and workload relief. It also shows how leaders can use AI as a tool while keeping the work human. Thanks for listening. Kirstin This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit leadthemachine.substack.com [https://leadthemachine.substack.com?utm_medium=podcast&utm_campaign=CTA_1]

Board-Ready AI: Leadership, Data, and Risk with Bill Walsh

Hi everyone, Bill Walsh, the CEO of Mediafly, has played a meaningful role in my leadership journey. I met him more than a decade ago when he served on the board at Valen Analytics, and he has been a steady mentor for me since then. Bill brings 30 years of executive leadership across enterprise software, analytics, logistics, and AI-driven solutions. He has led organizations through growth, product innovation, M&A, and cultural change. He has also served on public and private company boards, taught as an adjunct professor, and coached senior leaders. So what does “Board-Ready AI: Leadership, Data, and Risk” mean? Bill shares an adoption approach leaders can apply immediately: crawl, walk, run. Leaders learn the basics, use AI personally, and create a safe internal environment where teams can practice. Teams build confidence internally, and leaders use what they learn to shape customer-facing AI with clearer requirements and better discipline. Bill also makes the data requirement concrete. AI models depend on the quality of the underlying data. Leaders need accurate, consistent, complete, timely, and unified data to produce reliable outputs. Bill does not ask teams to make data perfect. But they do need to reduce errors, remove duplicates, clarify definitions, and connect silos in the places that matter most. He offers a practical lens for where to start: high value and low friction. He shares examples such as sales content personalization, predictive forecasting, churn analysis, recommendation engines, and knowledge and search. These use cases often deliver value without requiring a full rebuild of every system. We also talk about the board conversation. Boards prioritize risk and governance. Management teams prioritize innovation and speed. Bill advises leaders to bring transparency, explainability, and trusted expert input into board discussions so the company can move with discipline and direction. If you lead a team through AI adoption, or you support leaders who do, this episode gives you a clear framework you can use. Thanks as always for tuning in. Kirstin This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit leadthemachine.substack.com [https://leadthemachine.substack.com?utm_medium=podcast&utm_campaign=CTA_1]