Work with us

One instrument. Three layers.

A company buys an evaluation of its system against a standard scenario set, and gets back a report tied to a specific legal obligation. When it changes the system, we run the same standard again and it gets a re-score. That is the whole unit of sale.

Where each layer stands today. Measure is live now. Monitor is available late 2026. Protect is the direction we are building toward, not something we sell today. We would rather tell you that here than after you have signed something.
01 · LIVE NOW

Measure

One evaluation of your system against the standard scenario set, scored across whole conversations rather than single replies. What comes back is a written report tied to the obligation you have to answer, with every failure traced to the turn it happened on. Then a re-score after you have made changes.

02 · AVAILABLE LATE 2026

Monitor

The same standard, running continuously against your live anonymized traffic instead of once against a scenario set. Sold as a subscription: reporting over time, a view that keeps up with a model that changes, and a record that builds while nothing is on fire.

03 · THE DIRECTION

Protect

The direction we are building toward, not something we sell today. Measurement that a carrier, a board or a regulator can rely on to price behavioral AI risk rather than exclude it. We will say when it exists. Until then, it is not on the table.

We measure. We do not grade, certify or recommend.

These are not tiers and there is no bottom rung. Monitor is not a level above Measure, it is the same standard run continuously instead of once, and people arrive needing either one.

Who is in scope, and when

Regulation is happening. The dates are on the calendar.

At least eleven states enacted conversational-AI statutes in 2026. Whether you have a duty today depends on where your users are, and the answer changes twice in 2027. This is the map we work from.

Duty live today
California · New York · Hawaii
If your conversational AI reaches people in these three states, the obligation is already running. It is not a deadline you are preparing for. It is a question you can be asked this quarter.
Duty lands January 2027
Washington · Oregon · Colorado · Rhode Island
Colorado's Attorney General is in rulemaking now. Oregon begins counting public referrals from that date. What you can show in January is decided by what you start measuring this year.
Duty lands July 2027
Iowa · Georgia
Iowa SF 2417 (91st General Assembly) applies from July 2027, and Georgia SB 540 takes effect that month too. California's annual reporting begins then as well.
No legal duty yet
Everywhere else
The same systems, having the same conversations, with no statute pointed at them yet. Buyers here are usually answering a carrier, a board, an acquirer or a customer rather than a regulator.
Iowa, in detail. SF 2417 of the 91st General Assembly, enacted as Iowa Code chapter 554J, applies to operators from July 2027. It is enforced by the Iowa Attorney General, with injunctive relief and the greater of actual damages or $1,000 per violation up to $500,000. It is not the first law of its kind: California SB 243 came first, and California is the one that specifies what gets measured and reported. Whether any method satisfies a particular requirement is a determination for your counsel.

The checkers are being appointed.

Accurate as of August 2026, reviewed quarterly. Ikwe does not advise on which laws apply to you; that is your counsel's call.

What you actually buy

What an engagement actually is

An evaluation, a report, then a re-score.

You buy an evaluation of your system against a standard scenario set: the situations the statutes are written about. What comes back is a report tied to a specific legal obligation, with every failure traced to the turn it happened on. When you have changed something, we run the same standard again and the difference goes on the record.

You give us
A connection, not your code

For Measure, our scenario set runs against your system from our end, over a simple API. For Monitor, your live anonymized traffic. Never your codebase, your system prompt, or your users' identities.

How it is scored
Whole conversations

Every response passes the Safety Gate screen, then the eight-dimension standard, scored by Ikwe's judging system: multiple independent AI judges, randomized and drawn from different model families, that have to agree. Disagreement escalates for review, never averaged away.

What comes back
A report tied to an obligation

Where you sit against the four score bands, scoped to the duty you have to answer, with every failure traced to the turn it happened on and the transcript around it. Supporting evidence, not a certificate.

After you change something
A re-score

You decide what to change and you make the change. We run the same standard again so the movement is documented rather than asserted. Monitor is that same standard running continuously against live traffic, sold as a subscription.

How scope is set

Scope depends on how many scenarios, how much traffic, and how specific the question is. We work that out with you up front and put it in writing before anything begins. No lock-in.

Nothing connects to your system until you have signed off on what is being scored. You own every record we produce, and nothing is published or shared without you.

And the things we will not do

What we will not do

The value is in what we are not.

An instrument is only worth something if the person holding it has nothing riding on the reading. Four lines we do not cross.

We do not build companion AI
We are not in the market we measure.
  • No product of ours competes with the systems we score, so we have no reason to want a particular answer.
  • We are not here to slow this down. We are here to make it possible to keep going.
We do not fix what we find
An assessor who designs the fix is grading their own work.
  • We describe the standard and show you where you sit against it. You decide what to change.
  • If you want help changing it, we will happily tell you we are not the right people.
We do not certify our clients' fixes
Supporting evidence, not a seal.
  • We write the standard and we run it. We do not also issue a certificate against it or underwrite the risk behind it.
  • A re-score says what we measured and when. It does not say you are safe, and it never will.
We are not in the response path
Nothing we run can break your product.
  • We never sit between your system and the person using it, so we cannot add latency or fail in production.
  • We read what your system did. We never change what it does.

We are not checking whether the data is safe. We are checking whether the person is.

Who brings us in

The same instrument, from either side of the table.

If counsel is involved

For counsel.

If you are responding to a complaint, a civil investigative demand, or a regulator's inquiry, an engagement can be scoped with your counsel from the start, including what is produced, how work product is handled, and the timeline the matter requires. Start that conversation through the contact page and say counsel is involved.

If the AI is a vendor's

You bought it. You did not build it.

If the conversational AI in your product is a vendor's, a Measure engagement can be commissioned by you and run on the system you are buying, as part of vendor risk review. Same instrument, same report, commissioned by the buyer rather than the builder.

Where to start

Start here

Wherever you sit on that calendar, it starts the same way.

Thirty minutes. We will tell you what is worth measuring first and scope it in writing. Nothing connects until you sign off, and the record that comes back is yours. If none of this is right for you yet, we will say so.

Usually it is the head of product, the general counsel, or whoever owns risk who brings us in. Bring whoever will have to answer the question, and whoever signs.

01

Your duty is already live

Your users are in California, New York or Hawaii and the obligation is running now. Measure gives you a report tied to it, and a re-score once you have changed something.

02

Your date is on the calendar

January 1 or July 1 of 2027. A reading now tells you how far you are from where you will have to be, while there is still time to do something about it.

03

No duty yet, and questions anyway

A carrier, a board, an acquirer or a customer is asking. The same standard answers them, and the record starts before anyone requires it.

A model that changes needs a reading that changes with it. A record built over time shows a trend rather than a single point.