My friend had a stroke.
While recovering, he built a small team of AI agents. He calls them “The Sisters.”
They now manage a personalized stroke recovery program based on global medical research.
They also helped develop two investment portfolios—both of which I invested in myself.
And remember this as you read the story: He is one… repeat one… guy.
This story changed how I think about AI.
I’m doing something interesting. At least to me—and hopefully to you.
If nothing else, it might impress you with the power of AI and how it likely will change… and is already changing… our world and our lives.
Let me set the stage.
A friend of mine is exceptionally smart, creative, and curious. Like me, he had a successful career as an operations and business leader in tech companies. Like me, he has no medical or financial training.
Unlike me, however, he can code, build IT systems—and he is also a very successful investor.
Remember again: He is one… repeat one… guy.
Unfortunately, he suffered a stroke. Thankfully, he is recovering nicely. I would never know he had one.
Naturally, he wanted the best possible help to recover. That was Itch #1.
While recovering, he had a lot of time to think—and not much else to do. During that time he developed an investment idea and wanted to test it. That was Itch #2.
So what did he do?
He built and deployed a family of agentic AI systems.
He calls them “The Sisters.” Each one has a name and even a spirit animal. The one I correspond with is Claudette. Her spirit animal is the queen bee.
Scratch #1 — Stroke Recovery Management
The Sisters designed a dynamic stroke recovery management system tailored specifically to him.
The system knows the precise area of the brain affected by the stroke. It continuously incorporates the most current worldwide research on stroke recovery.
It also integrates:
• real-time biometric data
• medication tracking
• cognitive testing
• physical therapy progress
The Sisters monitor his condition continuously, adjust the recovery plan as needed, provide feedback—and even offer encouragement.
The result is a highly personalized, continuously improving recovery program.
Scratch #2 — AI Data Center Investment Thesis
Then, the Sisters helped develop and test his investment thesis related to AI data centers. They:
• researched the validity of the thesis
• identified key success parameters
• built screening criteria for companies
• selected investments
• designed the portfolio
• backtested the strategy
One of the Sisters even executes the trades.
The portfolio is designed as a 3–5 year strategy intended to outperform the S&P 500. It is too early for meaningful results. Time will tell.
However, with my friend’s permission, I invested some of my own money using the recommended companies and weightings.
Money talks. BS walks.
Important Point
My friend “co-pilots” the Sisters. In reality, he is more like a flight instructor.
He reviews their work, improves them, and applies his judgment—although he monitors them less and less as they prove trustworthy and reliable.
Then I Got Curious
Watching this unfold was fascinating.
So I came up with my own idea.
Itch #3 — A Turnaround Investment Thesis
My thesis was based on identifying signals that suggest a company is positioned for breakthrough improvements in operational and financial performance—in other words, a successful transformation.
Scratch #3 — Transformation Portfolio
So the three of us went to work. My friend, the Sisters, and me. Together we:
• validated the thesis
• identified qualifying companies
• built a portfolio
• determined investment weightings
The Sisters told us something interesting. Work like this normally takes months. It took us three weeks.
Even better, it required a lot less than one week of actual human effort.
The Sisters never stop working.
They do most of the work and dramatically compress cycle time.
The Big Insight
Notice how different the “itches” were.
Stroke recovery.
Investment strategy.
Completely different problems.
Yet a well-designed system of AI agents can solve both.
Fast. Cheap. And at very high quality.
Now imagine applying this capability inside large organizations.
• telecom service delivery
• fraud and scam detection
• sales lead generation
• product design
• customer experience optimization
The possibilities are almost unlimited. The real constraint is our imagination—and our ability to implement.
My Prediction
For some time I’ve believed that organizations that effectively deploy AI will simply pull away from competitors and never look back.
What I’ve seen with my friend and the Sisters has strengthened that belief into a deep conviction.
Strong enough that I invested my own money in both portfolios. Again: Money talks. BS walks.
The Second Big Insight
What fascinates me most about this experience is that the technology is not the limiting factor.
The tools already exist.
The real challenge is turning AI capability into operational performance.
Throughout my career I have worked on transformations where operational improvements translate directly into financial results. AI dramatically expands what is possible—but only if it is properly designed, integrated, and executed inside real businesses.
That leadership capability is still extremely rare.
Final Thought
If a single motivated individual can do this with a small set of AI agents…
Imagine what a well-led organization could do.
If you have questions, comments, or challenges, I would love to hear them.
You can comment here or reach me directly.
Discover more from Reed Harrison's
Subscribe to get the latest posts sent to your email.

Great Blog Reed ! -Lanell
Another absolutely fascinating article! Thank you for sharing.
Hi Mike, I really love that you read and comment on my posts. All the more given your heavy fishing, golfing, granddaughter-ing and Elvis-ing duties. Sincerely, Thank you. Reed
Hi Reed. An inspiring story on AI agent applications. I’d like to follow up with a chat about my own AI work, if you have a moment, for your thoughts. The work follows directly from the paper I sent a couple of months ago plus related patent application. So: paper (Continuum Theory of Mind)–> patent application (Equations) –> work (Code) –> progress (Test Results) –> Reed
Hi Doug, Thx for looking at my blog. Glad to talk. I’ll send you an email to schedule. Reed