ResourcesEvaluation
No. 02 / Updated May 2026
“We optimise for founders who are early, honest, and substantive, not founders who have already won somewhere else.”
Originn exists because the signal a founder needs in the first year is different from the signal a Series A investor needs in year three. Verified profile, real demand, a path to working capital, and a credible team, those four matter more than logo brands, prior exits, or how many followers you have on social.
This page is the canonical reference for what we look at. The dimensions are public. The exact weights, the scoring math, and the model configuration are internal, published, they would make the system gameable for everyone. We trade off some opacity there for usefulness everywhere else.
Dimensions are public. Weights are not. That is the trade.
Five dimensions, applied to every application. No exceptions, no shortcuts for connections.
Does your deck answer the questions it is asked? Are the twelve slides internally consistent, does the team you describe on slide ten match the credentials you cite on slide six, does the why-now on slide four support the market sizing on slide five? Reviewers can tell when a deck has been assembled from a template versus written by a founder who is actually building the thing. We weight written-by-a-founder higher.
Is the problem you describe a real one that real people have? Have you spoken to any of them, or is the problem statement a guess? Evidence does not need to be quantitative, three customer interviews quoted in plain language carry more weight than an unsourced "TAM = $50B" claim. We are checking whether you have done the founder work of going outside the building.
Why this team, for this problem, now? The credential we are looking for is the credential that matters in the specific wedge, not generic. A heart-rate-monitor founder with a sports-cardiology PhD scores high on this dimension; the same founder pitching a B2B procurement platform does not. We do not require a founder to be senior or pedigreed; we require them to be the right one for what they have picked.
Is there a real wedge that did not exist three years ago, and a real reason no one else has taken it? The two questions are linked, most strong differentiators come from a recent shift (technology, regulation, behaviour) that opens a wedge a smart team can occupy. "Everyone else's product is bad" is not a differentiator; "a recent change makes a new approach possible" is.
Does the architecture make sense? Does the build complexity match the team and timeline you have? Where applicable, we pattern-match against the shape of ventures that previously tried something similar and stalled, not to predict failure, but to surface the specific risks that the same wedge has historically produced. You are not penalised for being early in a hard space; you are flagged for under-estimating what the space asks of you.
Reviewers can tell when a deck has been assembled from a template versus written by a founder who is actually building the thing.
These are not requirements. They are the patterns that consistently correlate with applications that move forward.
These are not flags; they are reasons we decline. None are about being early.
Being early is not on this list. Being unincorporated is not on this list. Being a solo founder is not on this list. Being out of a non-IIT institute is not on this list. The disqualifiers are about honesty and substance, not pedigree.
Disqualifiers are about honesty and substance. Not pedigree.
A meaningful part of who Originn is for shows up in what we are deliberately blind to. None of the following affects your evaluation:
What you submit is what we read. Founders who do not have warm intros or media reach often have the sharpest problem statements, and those win on the dimensions above.
When your deck arrives, our internal evaluation system, Originn AI, does a first pass across the five dimensions above. It surfaces what looks strong, what looks weak, and what looks worth probing. The output is structured: a set of signals attached to specific parts of your deck, with reasoning.
A human reviewer then reads your deck end-to-end and forms an independent view. Where the system and the reviewer disagree, the reviewer wins. We do not auto-reject on a score, and we do not auto-approve on one either. The system is a research assistant, not a gatekeeper, the decision is always made by a person who is willing to defend it.
The system is a research assistant, not a gatekeeper.
Three honest reasons. We do not publish the score, the weights, or the internal scoring math.
Fairness across volume, we read every application personally. We do not have the bandwidth to write a one-page evaluation report for every applicant. What we can do, and what we do, is tell every declined founder the specific blocker that drove the decision, in one or two sentences.
Gameability, if we published exact weights, applicants would optimise for the score instead of for the substance. The score would stop being a useful signal. Withholding the weights is the price of keeping the system honest for the next applicant.
Intellectual property, our internal methodology, the model configuration, the prompt structure, and the way the five dimensions are combined is the thing that makes Originn AI useful at scale. We are not in the business of giving that to other platforms. We are happy to be opinionated about what we value; we are not in a position to publish how we measure it.
When you are readyThe next move
Free to apply. Reviewed by a real person inside seven business days. Decision is one of three, approved, needs changes, or declined with specific reasons.