IIRS: One Meeting Saved Per Reading, and the Same Verdict Next Month.

A pitch is a story told in the founder's order. Reading past it is real work, and a good partner does it. What a fund can do about it at scale is where the options run out. Put more partner hours on each company and the readings get deeper and still disagree. Write a shared checklist and it bends to whoever holds it, and to the week they hold it in. Send the work outside and you get one more opinion, in a format only the vendor uses. Hire more partners and each arrives with a bar of their own. Every option buys a better answer on one company. None makes it comparable to last quarter's. By the time anyone re-read a company, the bar had moved, and nobody had written down where it used to be.

IIRS produces three things about a company. A score on one scale. A verdict said plainly, from a fixed vocabulary: consistent, one caveat, fragile, earned, unproven, understated. A report written to be read aloud and survive the questions that follow. Underneath, signals move in different shapes and at their own pace, so the system models five behaviours rather than reading each alone: cyclical, uneven, stepwise, levelling, compounding. It takes the pitch apart and puts it back in the order the evidence suggests. Sometimes the story survives. Sometimes something better is underneath it. Scoring is evidence-linked: every element carries the material that produced it. The bar is chosen by what a company should look like by now, so a team of four and a team of two hundred are never measured against the same one. Ask it again next month and it returns the same answer, for the same reasons.

Refusal is the hardest behaviour to hold. The easy path is to reason from the categories around a missing disclosure and set the result on the page with the same confidence as a filed figure, with nothing in the report to distinguish the two. IIRS cannot take that path. The reason is structural rather than a matter of instruction. Because every element of a score is linked to the evidence that produced it, one with nothing behind it has nowhere to attach, and surfaces as a gap instead of a value. Where the gap matters, IIRS stops and puts a small set of precise questions to the founder. The same retained store does two other jobs. Each reading is kept with its reasons, so a new company is read beside the earlier ones it resembles. Every verdict also names a couple of things that should be true next, on a stated horizon. IIRS looks again when it arrives and marks each one held or not held.

The effects land on three clocks. This week: a decision that walks into IC with reasons that were not written to please anyone, and a founder who knows where they stand before the first meeting. This quarter: a single reliable reading saves one meeting, and one evaluation is read by many instead of the same story retold twelve times. Across a fund: every company measured the same way, so portfolio decisions can be compared, revisited and learned from. What we will not give you is a number. IIRS publishes no accuracy or outcome figures, and we are not inventing any here. Access is by request. The third order is a direction, not a result: it has not run long enough to show. What we do stand behind now: a no comes with reasons, in language a founder can act on. Nobody leaves with silence.