We build systems that have to show their work.

Kenshiki Labs builds evidence systems for decisions someone will be asked, later, to defend. Pulse is the first one aimed at people: proof that a real, present person is behind an account, at the moment it is opened and every time after.

Why we built this

Most systems that judge a person judge a photograph of a person, not the person. One reading is taken, and the reading is mistaken for the human. The error is cheap to make and lands hardest on the people who can least absorb it.

We decided that judging people without the obligation to show why was unacceptable to us. So we moved the proof out of a single snapshot and into something that can’t be forged cheaply: a real life, in motion, over time.

Inspection, or construction

The conviction comes from the quality tradition the founders trained in: the Deming and Toyota lineage that draws a hard line between inspecting a finished part and building quality into the process. Inspection is cheap and legible: measure the part at the end, accept or reject. But inspection can’t tell you why a part is out of spec, so it can’t tell a part that is bad from one that merely looks bad at the moment it was measured.

Most automated judgment is inspection applied to people. We build the other way: the proof is constructed from evidence before an answer is allowed to leave, and the final check merely confirms what was already built in.

Build to be examined. State uncertainty honestly. Prove before claiming.

What we believe, and what it costs

A value is only credible when it costs something. Ours do.

• When the evidence isn’t there, we return nothing. We accept fewer answers as the price of defensible ones.
• We corroborate the shape of a decision without surveilling its contents. Continuity can be proven without recording the merchant, the purchase, or the neighborhood. We leave predictive lift on the table rather than rebuild old exclusions with better tools.
• Every consequential answer reduces to claims that can be checked against evidence before it is emitted. Not tone, not topic. Claims.

Who it is for

We build for the desk where someone will be asked, later, to show why a decision was allowed. At that desk a fluent mistake can move money, shape care, or create legal risk, and it exists in lending, in government, in healthcare, and beyond. People are where we started, because proving a real person is the most exposed and most fixable instance of the problem.

The people

Kenshiki was founded by operators whose disciplines are where the conviction came from: statistical process control, Good Laboratory Practice, continental-scale data infrastructure, and regulated transaction systems all share one rule: you answer for the output under review later.

• Stephen Fishburn, Co-founder & CEO. Trained on statistical process control in the Deming and Juran lineage; later a Bell Labs researcher, a Microsoft Windows engineer, and the chief data scientist who built a pattern-of-life pipeline for 120 million mobile subscribers. That system is also why he won’t build one without limits. M.S. in computer science and mathematics, Tokyo Institute of Technology; bilingual in English and Japanese.
• Larry Signorile, Co-founder & CTO. Nearly four decades building systems that answer for their output later, from regulated defense work under Good Laboratory Practice to seventeen years in regulated consumer-data and verification, where he scaled a core transaction API from 100,000 to roughly seven million requests per hour and served as its security champion.