Friday, February 20, 2026
FRIDAY — AI FOR THE C SUITE
Read time: 11-12 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 coming Monday, February 23, 2026, I’m joined by Brent Orrell, Senior Fellow at the American Enterprise Institute and former Acting Assistant Secretary at the U.S. Department of Labor — where he oversaw a $10 billion workforce investment system. We dig into why the economy is shifting from doing to judging, and what a Procter & Gamble study found when AI-equipped teams vastly outperformed their peers in speed, quality, and breakthrough thinking, yet trusted their own work less. While you wait for that to drop Monday, check out this week’s episode to learn how to build cognitive depth for you and your team – a critical skill to avoid overload and burnout when using AI. Both are waiting for you wherever you get your podcasts.
Apple · Spotify · iHeart · Amazon · YouTube
Subscribe for free today on your listening platform of choice to ensure you never miss a beat. New episodes release every two weeks.
2. Algorithmic Musings: The Fence Is Now a Razor
|
TL;DR This one runs long. It needs to. Two major pieces about AI broke through to mainstream attention last week, and if your organization still hasn’t started building AI capability, the window for doing it on your own terms is closing fast. Read on for what to do about it. |
Everywhere I turn, I see AI fence-sitters. They’re patiently waiting for … something. The right moment. A clearer signal. A competitor to go first. A vendor to make it easy. (And yes, that includes all of you who like to tell me you’ve deployed Microsoft Copilot in your organization. That barely counts.)
I get it. The fence felt safe for a while. But that beautiful fence you’ve been sitting on? It just turned into a razor. And it’s time to move before it cuts you.
If that describes you and your organization, consider this one of your last gentle nudges to get moving.
Two Pieces You Need to Read
Last week, two pieces of writing about AI broke through to mainstream attention in a way I haven’t seen before.
The first is Matt Shumer’s blog post, “Something Big Is Happening,” which has now been viewed over 80 million times. Shumer is an AI startup founder writing an open letter to his non-tech friends and family, and his core message is blunt: the gap between what AI insiders are seeing and what the general public understands has become dangerously wide.
The second is Josh Tyrangiel’s cover story in The Atlantic, “America Isn’t Ready for What AI Will Do to Jobs.” Built on dozens of interviews with economists, workers, and tech leaders, it examines how AI is already reshaping the labor market for entry-level knowledge workers and asks the question nobody in politics wants to touch: what happens when this technology matures and we’ve done nothing to prepare?
Neither piece surprised me. We’ve been discussing these ideas here and with our clients for years. What did get my attention was the scale of public reaction. The conversation has shifted. The question is no longer “Will AI change things?” It’s “Why haven’t you started preparing?”
And that question lands differently when you work with middle-market organizations every day, which I do. The large enterprises have dedicated AI teams. The startups were born in this environment. But most middle-market companies are still watching from the sidelines, telling themselves they’ll figure it out when the technology “matures.” The problem with that logic: the technology is maturing while you wait. And it’s not waiting for you to catch up.
Build the Muscle Memory Now
If your organization doesn’t start building AI muscle memory today, you will get absolutely flattened when the weight doubles. Here’s what I mean.
Muscle memory is what happens when you practice something enough that it becomes second nature. Athletes build it. Musicians build it. And organizations need to build it with AI. Not because the tools you use today will be the tools you use in two years, but because the organizational capacity to adopt, adapt, and integrate AI is the real competitive advantage. You build that capacity by trying things, screwing some of them up, and learning from both. It doesn’t show up on its own.
When you wait, here’s what you’re really signing up for:
Your competitors will outrun you. While you’re still debating whether to form a committee, they’ll be on their second or third round of implementation. Their margins will tighten. Yours won’t.
Your best people will leave. They’ll hear from friends at other companies who no longer spend their days on manual double data entry or other tasks that AI handles effortlessly. They’ll wonder why they’re still doing it the hard way. And then they’ll stop wondering.
The inertia will harden. The organizational resistance keeping you from moving today doesn’t soften with time. It calcifies. And eventually, the internal champions who would have driven the initiative will have moved on to organizations that actually let them lead.
Your “Competition” Is About to Look Very Different
Here’s something that neither Shumer’s post nor the Atlantic piece fully addresses, but that I think about constantly when I’m working with middle-market leaders: your competitive landscape is about to change in ways you haven’t imagined.
For the past several decades, competition in knowledge work has been defined by scale. You competed against firms of roughly similar size, with similar overhead, similar talent pipelines, and similar go-to-market capabilities. Barriers to entry were real. Building a competing firm required capital, people, and time.
Those barriers are dissolving.
A single individual with some technical skills, or even just a friend who has them, can now build capabilities that would have required a team of ten just two years ago. They can generate professional-grade marketing. They can build and deploy software. They can analyze data, draft proposals, manage client communications, and operate like a mid-sized firm used to. Their overhead? A laptop and a few software subscriptions.
This isn’t theoretical. It’s happening right now, and it will accelerate. Your next competitor might not look anything like your current ones. They might be a two-person operation that punches like they have fifty people. And they will come for your clients with lower prices, faster turnaround, and the kind of agility that a larger organization struggles to match.
The only way to counter that is to pick up the same tools yourself. Not someday. Now.
So What Should You Actually Do?
I know you didn’t come here for abstract warnings. You came for the recipe. Here’s where to start:
One. Create AI policy. You probably don’t have one. Your employees are likely already using AI tools in ways you don’t know about, with data you haven’t thought to protect. Get something on paper. It doesn’t need to be perfect. It needs to exist.
Two. Buy team licenses. Get at least five people (more if you can) set up with Claude or ChatGPT. Not free accounts. The paid tiers. The capabilities gap between free and paid is significant, and you can’t evaluate what this technology can really do for your organization if you’re kicking the tires on the economy model.
Three. Form an AI exploratory committee. Pick a small group of curious, motivated people from across the organization. Give them a mandate: find out where AI can help us. Not a massive strategic initiative. A scouting party.
Four. Set expectations and deadlines. Tell that committee you want a preliminary report in 60 to 90 days. What did they try? What worked? What surprised them? What do they need? Without a timeline, exploration turns into tourism.
Five. Start using it yourself. If you’re a leader reading this and you haven’t personally spent meaningful time with these tools, that’s a problem. You can’t lead an AI transformation you don’t understand. Set aside 30 minutes a day. Ask it to help you with something real. Draft a client email. Analyze a report. Summarize a document. Experience it firsthand.
Six. Prepare for the next step. After three to six months of genuine organizational exploration, you’ll know enough to make informed decisions about where to invest more deeply. That’s when you bring in outside help. Not before. You’ll be a smarter buyer of those services if you’ve done the preliminary work yourself.
The Window Is Closing
I’ll leave you with this. There’s a moment in every wave of significant change when preparation shifts from “smart and proactive” to “desperate and expensive.” We’re still in the first category, but not for much longer.
Shumer compared this moment to February 2020. The analogy isn’t perfect, but what is similar is the gap between what the people closest to the situation understand and what everyone else is willing to accept. With COVID, that gap closed in about six weeks. The difference this time? You’ve got people waving the flag before the world turns upside down. You have runway.
But runway has an end.
If you’re ready to start building AI muscle memory in your organization and want some help figuring out where to begin, drop me a line. Let’s figure it out together.
3. Research Roundup: What the Data Tells Us
Prompt Repetition: The Zero-Cost AI Accuracy Fix You’re Not Using
Google researchers just handed every AI-deploying company a free performance upgrade, and it’s almost embarrassingly simple: repeat your prompt before asking for a response.
The numbers that matter: Tested across seven major AI models (Google, OpenAI, Anthropic, DeepSeek), prompt repetition produced statistically significant accuracy gains in 47 out of 70 test cases—with zero cases where it hurt performance. On position-sensitive tasks like extracting data from structured lists, one model jumped from 21% to 97% accuracy. The technique adds no latency to response generation and requires no retraining or new tools.
What this means for your Monday morning: If your team is using AI for classification, data extraction, form processing, or any task that doesn’t involve step-by-step reasoning, you can improve accuracy today by duplicating the prompt in your instructions. It’s a five-minute configuration change to existing workflows.
The catch: You’re doubling your input tokens, which means higher input costs. For high-value tasks where accuracy matters—invoice processing, contract analysis, customer data extraction—the math works easily. For high-volume, low-stakes tasks, run the numbers first.
Action item: Have your AI lead audit your non-reasoning workflows this week. Pick the three where accuracy matters most, implement prompt repetition, and measure the before-and-after. You’ll have hard data on whether to roll it out broadly within days.
Read our full analysis of this and all other analyzed research papers at AI for the C Suite.
4. Radar Hits: What’s Worth Your Attention
AI coding tools for knowledge work: what executives need to know. If your team thinks tools like Claude Code are only for developers, they’re wrong. MIT Sloan lays out how agentic coding tools handle competitive intel refreshes, campaign versioning, and due diligence reviews by working directly with your files. The key difference from chatbots: built-in memory, automation, and parallel execution. Worth a pilot outside engineering.
Using AI reduces new skill formation, study finds. Programmers who leaned on AI scored 17% lower on skill assessments than those who didn’t. The fix: require AI to explain its work, not just generate output. If you’re rolling out AI tools to junior staff, build “learn first, automate second” into your training protocols.
Anthropic publishes the prompt injection data enterprise security teams have been asking for. Anthropic broke out prompt injection attack success rates by agent surface, something OpenAI and Google haven’t matched. If you’re evaluating AI vendors for anything that touches sensitive data, this is the new procurement baseline. Ask every vendor on your shortlist for equivalent numbers.
AI research studies can suffer from outdated models. Studies making rounds about AI limitations are often tested on models that are already two generations old. Before you pump the brakes on an AI initiative because of a headline, check which model was tested. Last year’s benchmarks don’t apply to this year’s tools.
5. Elevate Your Leadership with AI for the C Suite
I spent most of this week’s newsletter telling you to get off the fence. So let me make it easy.
If you’ve been reading this newsletter and nodding along but haven’t taken a concrete step yet, here’s your assignment: pick one item from the action list above and do it before next Friday. Just one. Create the policy. Buy the licenses. Form the committee. Start using the tools yourself. Then reply to this email and tell me which one you chose. I read every response, and I’ll point you in the right direction.
And if you’re past the starting line and ready for the next level, let’s talk. I’m booking Q2 strategy sessions now, and I’d rather help you build this while it’s still proactive than after it becomes urgent.
See you next week.
Stay safe. Stay healthy. Be strong. Lead well.
Chad