The lens
AI is everywhere. Moats are not. We isolate one lens — who is turning intelligence into a durable operating advantage — then score it. Three pillars, 0–5 each, total 0–15. The hard fence on every pillar is 2 vs 3.
2 = operational hygiene (same business, run cleaner). 3 = could move business model or market position. If every competitor bought the same Copilot seats, would this company still look different? If no → 2.
How much of the chain has operating use. One interesting place can be a 3 - especially how they go to market. Breadth still matters for a 4 or 5.
Did it land. 2 = named in a 10-K or earnings, tied to a function. 3 = you can see it running. 5 is a KPI on the intelligence function — not robot or device volume.
A 3 or above on a pillar means we found a genuine project and a result we can point to — in the product, the operations, R&D, or how they go to market. 2 = they are talking about it. 1 = mentioned, still to be determined. 0 = nothing. A total of 9 (three 3s) means they are doing something worth a closer look - it might move the needle for shareholders. One place on the chain is enough; how they go to market is more interesting than ops or sales noise. We do not hand out 4s for robot placements or sensor volume. A live product that was a 3 stays a 3 until they sunset it.
We are early in the cycle. A company can be a real bet at a 3 or 4. None of the first Health Care Equipment ten earned a 5.
The map starts from the S&P 1500 — the S&P 500, MidCap 400, and SmallCap 600 — as held by the iShares IVV, IJH, and IJR funds. That is a free proxy, not a licensed S&P list. We then drop the tip of the spear (chip and cloud vendors, most of IT) and scan traditional operators. We do not score all 1,500. The product is a short list inside that pond.
We do not read everything about every company. A name has to clear a hurdle before we spend the next kind of attention. Most names stop at the first pass. We are looking for a signal, not writing up the news. What the company says is one input. What other people report is another.
Pass 1
The earnings release, the transcript, the investor deck, and a look at what others are saying about the company and those initiatives. The annual report is the baseline. It is not the last word.
Hurdle: a signal. A use they did not put in the release, a customer, a result, or a real disagreement. A slogan or a repeated announcement is not enough.
Pass 2
Only when the first pass found something of interest. We look harder at that one thing, to verify it, test the claim, and add context.
Hurdle: the extra look confirms the signal, knocks it down, or adds a fact we did not have. A repost of the announcement does not count. This pass can change the score.
Pass 3
When the signal is live, we hand it to a person. Test the product, call someone, or ask investor relations. That work is human.
What comes back can move the score. Until then, the number on the site stays where it was last accepted.