Adopt AI coding agents without
failing your next exam.
MGAs and program administrators, regional and farm mutuals, public-entity risk pools, and claims TPAs must build custom software without an engineering bench: submission intake, raters, bordereaux, portals, statutory reporting. The enterprise vendors’ AI factories disqualify this tier. We built for it. Agents are the way out of the backlog, and every board knows it. But every compliance regime rests on three questions: who changed the system, who reviewed it, who approved it. And the AI bulletins have added a fourth: why did the system decide that?
When agents join the team,
the attestation chain breaks.
“A qualified person wrote it, a second person reviewed it” — stops being true the day an agent writes the code. Haltere repairs the chain: every change carries a human name, a recorded review, and a durable approval. By construction, not by policy memo.
Three regimes. One record.
The examiners aren’t hypothetical. These frameworks already govern how your software changes. And 24 states plus the District of Columbia have adopted the NAIC model bulletin on insurers’ use of AI, per the NAIC implementation map dated April 1, 2026. What the examiner will ask for next
“SOX for insurers.”
IT-control evidence per system, per change. Today that evidence is assembled by hand, after the fact. On Haltere it’s generated by the system, not compiled by a team.
The change record, with names.
Examiners ask for the change record. They expect a human name on it. The receipt keeps one there: actor, reviewer, approver, on every change an agent touches.
The AI questions are now in writing.
NYDFS’s AI guidance and amended Part 500 ask for application-security and audit-trail exhibits that map to records that already exist, because the system can’t run without producing them.
One receipt per change. One query per audit.
Every step the AI took,
and who signed off.
A single rating change: retrieved, drafted, checked, approved, sealed. Each step on the record, and every rating change tied to its filing. We show the highest-stakes example on purpose; your pilot starts wherever you choose: a portal, a report, a workbook. The receipt’s shape is identical.
Fixture data. A value outside the filed band fails at the gate. No human is ever asked. The receipt is the difference between answering the examiner and reconstructing for them. Download the sample evidence pack (PDF, fixture data)
Bad changes don’t get reviewed. They get blocked. Merged: 0 lines, logged, receipted.
The exam won’t only ask
how it was built.
It will ask what it did. Your software applies rates, routes claims, flags files. And when the letter arrives about one decision from March, reconstruction is either a query or a project. On Haltere the operation already wrote its decision receipt: the actor and their authority, the model and prompt that ran, the inputs verified back to source, the checks that passed, the trace.
The bulletins tell insurers to govern their AI, and to answer for the third parties who act on their behalf. They offer no mechanism to comply. The receipt is the mechanism. A query, not a project.
Carriers, and the vendors
who serve fifty at a time.
The governed first agentic project.
Mutuals, regional P&C, workers’ comp funds: the carriers whose paper the programs run on. One backlog item you already want, delivered governed in 4–8 weeks, with the evidence pack that lets you show the board AI progress and the examiners the receipts. The first cohort of five founding design partners is filling now, and carriers are welcome in it.
The proof layer under your AI hub.
Every policy-admin vendor is shipping an AI hub. The hub decides. Increasingly, it builds. And your carriers will be asked to prove both: how the software changed, and why it decided. Your agents, your models, your hub; our map, gates, and receipts. The hub ships faster, the evidence writes itself, and delivery hours stop eating margin.
The governed first agentic project.
One backlog item. Fully governed. Evidence included.
Submission intake and clearance. The Excel rater that should be an auditable system. The bordereaux you still assemble by hand. The portal rebuild that never got staffed. Scoped together before anything starts.
Scoped to the project and quoted in the first conversation. Agreed before anything starts. No time-and-materials meter running.
From scoping to production-acceptable delivery, live in your environment.
Two deliverables, always paired: the working software, and the evidence pack. Show the board AI progress; show the examiners the receipts.
The process guarantee: your team reviews and approves every change. Nothing merges ungoverned. Zero unauthorized code reaches production.
Two ways to run it: we operate the line for teams with no engineering bench, or your team operates it and we train your operator. The software, the map, and the receipts are yours in either mode.
You own it. The software and the logic built on top are yours. After the pilot, the line runs as governed operations: we operate it for you, or we train your operator and your own team builds the next ten projects the same way. You walk at renewal with everything. The pilot is how you adopt agents; governed operations is how you keep them governed.
Mapped to the exams
you already sit.
Generated by the system, not compiled by a team. The technical file becomes an export.
The long tail below the tier-1s
has no bench of engineers.
U.S. insurers, plus the MGAs, program administrators, and TPAs who write and service on their paper. Most must build custom software with teams they can’t grow.
state regulators. Custom software is mandatory; ungoverned software is a finding.
states, plus the District of Columbia, have adopted the NAIC model bulletin on insurers’ use of AI (NAIC map, April 1, 2026). The questions are already on the exam.
The wedge is insurance: MGAs and program administrators, regional and farm mutuals, public-entity risk pools, claims TPAs. The tier the enterprise vendors’ AI factories disqualify, given world-class technology without a full engineering-team price tag. The market is the regulated mid-market: community banks and credit unions, healthcare payers, government suites, FDA-regulated quality systems. Everywhere organizations must build software, can’t hire engineers, and answer to an examiner. Governance is the purchase driver here, not a tax on the sale.
Bring us a program with a backlog,
a carrier, or a vendor with an AI hub.
We’ll bring the receipts. Scoping is a conversation, the fee is fixed before anything starts, and the first deliverable review is on your calendar before we write a line.
No newsletter. No drip sequence. Every request gets a real reply within one business day — from someone who can answer it.
