We Started by Measuring Founders the Usual Way. The Usual Way Couldn't Tell Them Apart.
Kaleion Labs exists to give every founder the same quality of honest read, whoever they know — and to build the evidence base that shows what actually helps them move.
Where this started
A Result Nobody Was Looking For.
The company began with a doctoral study of 1,439 entrepreneurs. On every self-report readiness item, the averages fell between 4.0 and 4.4 out of 5, and 78–91% of respondents chose 4 or 5. The survey could not tell a founder with three customers from a founder with an idea.
That finding points one way. If you want to know where someone actually stands, you have to look at what they have done — and somebody has to build the thing that can do that at scale, carefully, and refuse to use it against them.
On a scale from 1 to 5, all five self-report item means fell between 4.0 and 4.4. The range from 1 to 4 is empty. 78–91% of respondents chose 4 or 5 on each item, n = 1,439.
78–91% of respondents chose 4 or 5 on each item · n = 1,439
How we work
Six Rules We Hold Ourselves To.
Evidence before assertion
Every figure carries its sample size. Every claim has a test that could fail.
Facts about the world, never about the person
"Nobody has paid yet" can be acted on. "You're not ready" can only be accepted or resented.
One move, not a curriculum
The diagnosis may be broad. The next action is always narrow.
Development, never selection
No ranking, no gatekeeping, no export that orders people. Ever.
People stay in the loop
Human adjudication is permanent. It is how expert judgment enters the system.
Publish either way
Pass conditions are set before a test runs, and results are published whichever way they fall.
The vision
A Record of What Actually Helps Founders Move — Built One Read at a Time.
Today, nobody in this field can say with evidence which kind of support moves which founder, at which point, in which circumstances. Programs spend on curricula, mentors and events because those are what can be delivered and counted.
Our north star is simple: know what is true, find the binding constraint, create the next piece of real-world evidence, and see what changes. Repeated across many founders and many programs, that loop builds a longitudinal record linking a measured position, the move prescribed, and what happened next.
If the evidence holds, that record lets programs route the right support to the right founder at the right time — and lets a founder anywhere get a read as honest as one they would get from someone who has done it before. We will make each of those claims only when the gate for it has been passed.
How we earn each claim
Each Stage Sells Only What the Last One Proved.
A ladder of six stages, from stage 0, descriptive, at the bottom — where Kaleion Labs is now — up to stage 5, long-run outcomes, which may never be built.
5 · Long-run outcomesAn open governance question. It may never be built.
4 · What helps whomOnly after the gates pass
3 · Near-term predictionOnly after the gates pass
2 · SensitivityRequires a comparison group
1 · ReliabilityNext — tested in the first pilot
0 · DescriptiveWe are here
Stage
The question
Status
0 · Descriptive
What is true now?
We are here
1 · Reliability
Would trained reviewers reach the same conclusion from the same evidence?
Next — tested in the first pilot
2 · Sensitivity
Does the read change when reality changes, and not just because people learned the test?
Requires a comparison group
3 · Near-term prediction
Given this position, how likely is a proof event or a stall in the next weeks?
Only after the gates pass
4 · What helps whom
Which move is more likely to work for someone in this position and context?
Only after the gates pass
5 · Long-run outcomes
Which patterns predict a durable business?
An open governance question. It may never be built.
Stage 5 is the capability most likely to be repurposed for selection, whatever a contract says. The honest position may be that we never build it. That decision belongs to the founder and, later, the board.
What compounds
Models Are Replaceable. The Evidence System Is Not.
Better foundation models make this more valuable rather than less: the frame and the adjudicated record are what turn a general capability into a specific instrument.
The coding frame
Anchors written as observable behavior, versioned like software.
The adjudicated record
Expert judgment on a decision nobody has recorded at volume before.
The longitudinal record
Founders read over months, with every past read reconstructable.
The prescription record
Position, constraint, move, format and outcome, joined.
Where we are
Pre-Revenue, Pre-Pilot, and Saying So.
We have no customers, no completed pilot and no reliability data yet. We have the founding research, a specified method, published gates and a sequenced build. The full list of what exists and what we are not claiming is on the homepage.
Jorge's doctoral research surveyed 1,439 entrepreneurs and produced the ceiling-effect finding the company is built on. Based in Las Vegas, he works with founders and the programs that serve them across the U.S.–Mexico corridor, with a focus on first-generation, bootstrapped and owner-operator founders, including Spanish-speaking founders.