Friday, April 24, 2026
FRIDAY – AI FOR THE C SUITE
Read time: 6-7 min · Read online
Hi, it’s Chad. Every Friday, I serve as your AI guide to help you navigate a rapidly evolving landscape, discern signals from noise and transform cutting-edge insights into practical leadership wisdom. Here’s what you need to know:
1. Sound Waves: Podcast Highlights
This past Monday, I was joined by Ondar Tarlow, a Diamond Award-winning and Top 100 Marketer who served as CMO at two major financial institutions and now consults with mid-market companies on AI adoption. Ondar believes that AI literacy comes from doing, not watching. He lays out a minimum viable plan and explains why treating ChatGPT like a search engine means you’re not even scratching the surface of what these tools can do for your business. Hit one of the below links to check it out:
Apple · Spotify · iHeart · Amazon · YouTube
Subscribe for free today on your listening platform of choice to ensure you never miss a beat.
2. Algorithmic Musings: A Consultant Looks at Forty (Months)
The late, great poet sage of the late 20th/early 21st century once famously said “I took off for a weekend last month just to try and recall the whole year.” (As the kids like to say, if you know, you know.)
Anyhoo, I also recently took a weekend off with the goal of trying to recall (and understand) the last three and a half years of my journey with AI.
The method wasn’t journaling. It was a structured self-interview: nineteen questions furnished by our AI developer, designed to surface how I actually prompt and interact with AI, not how I think I do. The output was a digital twin specification, a document describing my methodology for working with AI (and how I teach others) well enough that another system could eventually behave like me.
What surprised me? The methodology was already there. I just couldn’t see the whole thing from the inside.
Three and a half years of daily use had produced a working system: mental models, anti-patterns I instinctively avoid, metaphors I lean on, project architectures I’d built without naming them, verification moves I run almost automatically. Some of it I was aware of. A great deal had become invisible to me. Volume and muscle memory had hidden the design.
You don’t need to build a twin to get the same benefit. You need the move that makes twin-building possible: structured self-interrogation about a domain you’ve logged enough hours in to have a real practice but haven’t yet articulated.
Four suggestions if you want to try this on your own work.
Pick a domain where volume exceeds vocabulary. The value of this exercise scales with how much tacit practice you’ve built. Interview yourself on the thing you do a lot but have never had to teach.
Use someone else’s questions. Your own questions reinforce what you already think you do. Outside questions surface what you’ve actually built. If you can’t find a questionnaire, hand the task to an AI and instruct it to interview you with the discipline of a methodology researcher.
Use AI as your interviewer. Either within a chat or as a structured project, have AI guide you through the questions. Provide samples of your work and thinking for analysis. Don’t allow it to “give you answers” but do lean on it for large volume pattern analysis, distillation and summarization of your own thoughts and as an exercise partner.
Write down the answers, then re-read them as if a stranger wrote them. That’s where the pattern recognition happens. Your methodology emerges from the aggregate, not from any single answer.
The deliverable at the end doesn’t have to be a digital twin. It can be a one-page summary of your working principles, a list of anti-patterns you’ve learned to avoid, a named set of moves you make without thinking. What matters is the externalization. Methodology you haven’t written down erodes. Methodology you’ve written down compounds.
The weekend paid for itself in the first hour. Come Monday (and yes, that was Buffett up top), I had something I didn’t have on Friday: a written map of my own methodology and a roadmap for our AI developer to code my mental model and operational framework… stay tuned for some exciting news on that subject later this summer.
If you’ve got a domain where volume exceeds vocabulary and want help running the interview on yourself, drop me a line.
3. Research Roundup: What the Data Tells Us
Leaders often won’t commit to a call until the evidence is overwhelming and the downside is somebody else’s. New research from GE Aerospace, published in IEEE this month, finally gives that problem a name and number.
The numbers that matter: In organizational life, blame outweighs credit by roughly four to one. That asymmetry is the engine behind hedged recommendations, overengineered proposals, and opportunities that pass while your team gathers more data. Traditional tools help with complexity, but half of what executives navigate (volatility and ambiguity) resists upfront preparation entirely.
What this means for your Monday morning: The fix isn’t better analytics. It’s treating decisions themselves as assets worth capturing. When criteria, alternatives, and known unknowns get recorded at decision time, and agentic AI monitors them against shifting conditions, leaders get “qualified accountability,” meaning protection from being second-guessed on knowledge they couldn’t have had.
The catch: This is a conceptual framework, not a validated product. The technology exists, but the cultural prerequisite has to come first: guaranteeing good-faith decisions won’t be punished later. Without that, your team will keep hedging their records.
Action item: Pick one cross-boundary decision category (supplier selection or engineering tradeoffs work well) and require a one-page decision record capturing assumptions and unknowns. Middle-market firms can stand this up faster than enterprises stuck with legacy systems.
Read our full analysis of this research at AI for the C Suite.
4. Radar Hits: What’s Worth Your Attention
Mustafa Suleyman: AI development won’t hit a wall anytime soon—here’s why. Microsoft AI’s CEO forecasts another 1,000x jump in effective compute by end of 2028 and a shift from chatbots to agents that handle weeks-long projects autonomously. Consider the source, but the capital and gigawatt-scale buildouts are public record. If your three-year plan assumes AI capability plateaus, rebuild it. Ask your leadership team which cognitive roles still look defensible when AI teammates work for days without supervision.
OpenAI open-sourced a tool that scrubs secrets before ChatGPT ever sees them. The 1.5-billion-parameter Privacy Filter runs locally, masks PII and API keys with 96% accuracy on standard benchmarks, and is free under Apache 2.0. Your employees are already pasting client data into chatbots. This is a cheap governance layer between them and the cloud, worth piloting alongside your existing DLP stack before your next audit.
Who owns AI outcomes? Fortune AIQ’s new survey has uncomfortable answers. 59% of executives name the CEO as ultimately accountable for AI results, yet only 16% say AI responsibilities are clearly defined and 60% say AI performance isn’t tied to compensation or promotion. If you’re the CEO, you already own this whether you’ve formalized it or not. Draft an AI governance charter now that names owners for specific outcome classes, before your first incident forces one on you.
5. Elevate Your Leadership with AI for the C Suite
Know an executive worth a conversation?
AI for the C Suite has crossed 60+ episodes and a library of operators, researchers, and applied practitioners. I’m scoping Q4 2026 guests now: mid-market CEOs, operators running real AI implementations, or contrarians with a thesis worth stress-testing.
Know someone (or you)? Send a short pitch. One paragraph on the angle is enough.
And if you know a mid-market leader who should be reading this, forward it along.
Stay safe. Stay healthy. Be strong. Lead well.
Chad