AI Readiness and Responsible Decision Advisory for MGAs and MGUs
AI Governance Advisors (AIGA) is a senior-led, platform-neutral advisory that helps U.S. managing general agents (MGAs), managing general underwriters (MGUs) and other delegated-authority intermediaries govern their use of AI, starting with the next AI vendor decision.
Inventory. Classify. Decide. Evidence.
AIGA helps you review AI vendors before you sign, decide which underwriting, claims and fraud uses can proceed and on what conditions, and keep the evidence your carriers, capacity providers, board and examiners will ask for.
We work with CROs, CCOs, and heads of underwriting, claims, operations, vendor risk and technology at MGAs and MGUs. AI is reaching MGAs mostly through vendors and embedded features, often faster than ownership, approval and records can keep up. We close that gap without slowing the business down.
We work with the tools and platforms you already have. No software purchase required.
Principals scope and deliver every engagement themselves.
Every decision and every artifact stays with your team.
MGAs come to us when they have a specific decision and a date attached.
A new or renewing vendor for submission intake, underwriting support, claims triage or fraud scoring.
A request that arrives at onboarding, renewal or a program review.
A board directive or a request for executive visibility.
A deadline set by someone outside the MGA.
A single use case that needs a defensible yes, no or "yes, if."
These are practical diligence questions, not legal advice. Contract terms belong with your counsel.
An unanswered question does not end the review. Record it, then decide whether to condition it, fix it before signature, or escalate it.
The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers (adopted December 4, 2023) says an insurer's AI program should address third-party AI systems and data, including due diligence and, where appropriate and available, contract terms on audit rights and regulator cooperation. This checklist turns that expectation into questions an MGA can put to a vendor. Last reviewed: October 8, 2026. Source: NAIC Model Bulletin, Section 3, guideline 4.0.
The MGA AI Vendor & Procurement Risk Review Sprint runs this review for you as a fixed-scope engagement. It delivers a vendor register entry, a risk tier, a decision memo, approval conditions and an oversight calendar.
Our four-step framework — Inventory > Classify > Decide > Evidence — gives every MGA a structured, defensible path through AI governance.
Illustrative example only. This is a fictional vendor and scenario showing the format. It is not client work and not an outcome AIGA has delivered.
MGA AI Vendor & Procurement Risk Review Sprint
A bounded review of one or more AI vendors: vendor register, risk tier, decision memo, conditions and oversight calendar
MGA Governance Launch & Evidence Pack
Your baseline operating model: AI inventory, RACI, intake and approval workflow, control matrix, evidence index and executive pack
MGA Control & Examination Readiness Sprint
A readiness matrix, evidence index, issue log and remediation roadmap focused on a specific carrier, audit or examination request
Managed AI Governance
Inventory refresh, triage of new uses and vendors, vendor monitoring, KRI reporting, and committee and board support. Available once a baseline is in place and you have a named internal owner
The National Association of Insurance Commissioners (NAIC) adopted its Model Bulletin on the Use of Artificial Intelligence Systems by Insurers on December 4, 2023. The bulletin is addressed to insurers and sets expectations for how insurers govern AI used by or on behalf of the insurer, including due diligence on third-party AI systems and data.
States adopt the bulletin one at a time. The NAIC implementation map (status as of August 31, 2026) lists 25 states and the District of Columbia as adopters, and California, Colorado, New York and Texas as having their own insurance-specific AI guidance or regulation. Because MGAs act on behalf of carriers and also buy AI from vendors, carriers may ask their MGAs for related evidence. How these rules apply to your MGA is a question for your counsel. We help you build the records that answer the question.
AIGA is a senior-led U.S. advisory that helps insurance MGAs, MGUs and other delegated-authority intermediaries govern their use of AI. Work starts with a specific decision — a new or renewing AI vendor, a carrier request, or a high-impact use case — and ends with records the MGA keeps. AIGA is platform-neutral and does not provide legal advice, audit, certification, actuarial work or model validation.
Review the vendor before signature, at a depth that matches how close its AI sits to underwriting decisions made under your carrier's authority. First, pin down whether the tool drafts, recommends or decides. Then ask written questions, and request evidence, in nine areas: purpose and limits; data and data rights; accuracy and unfair-discrimination testing; explainability and records; human review and override; change notice; incidents, audit rights and regulator cooperation; subcontractors and model providers; contract terms. Close with a recorded decision, named conditions and a reassessment date. AI Governance Advisors (AIGA) runs this as the MGA AI Vendor & Procurement Risk Review Sprint, a fixed-scope engagement that delivers a vendor register entry, a risk tier, a decision memo, approval conditions and an oversight calendar. Contract language stays with your counsel. Model validation, fairness testing and security testing stay with qualified specialists.
The bulletin is addressed to insurers, not directly to MGAs, but carriers may pass similar expectations to the MGAs that underwrite or handle claims for them. Whether and how a given state's bulletin reaches your MGA is a legal question for your counsel. AIGA helps you build the inventory, decisions and evidence that conversation needs.
Send a short, current evidence pack — not a policy document alone. It should show which AI tools touch underwriting, claims or fraud decisions, how each is risk-tiered, who approved it and under what conditions, where human review sits, and which items are still open. AIGA's MGA Governance Launch & Evidence Pack builds that baseline.
Every engagement starts with one vendor, use case or external request and a date that matters — a contract signature, a renewal, a carrier deadline or a board meeting. After a short scoping call, AIGA proposes a fixed-scope engagement in writing. MGAs usually begin with the MGA AI Vendor & Procurement Risk Review Sprint, then move to the MGA Governance Launch & Evidence Pack.
No. AIGA does not give legal or regulatory opinions, and it does not audit, certify, provide assurance, perform cybersecurity testing, do actuarial work or validate models. AIGA helps MGAs make accountable AI decisions and keep the evidence behind them.
AI Governance Advisors (AIGA) helps MGAs and MGUs build that evidence. Carriers usually want to see which AI tools and vendors touch underwriting, claims and fraud decisions, how each was reviewed and approved, and the conditions on each. They also look for where people review and override outputs, how vendors are monitored, and what is still open. If the question is about one vendor, the MGA AI Vendor & Procurement Risk Review Sprint produces that vendor's decision record. If it is about your program as a whole, the MGA Governance Launch & Evidence Pack builds the inventory, approval workflow, control matrix and evidence index. AIGA does not attest to or certify your program. The records show what you reviewed, what you decided and what you control.
No. A SOC 2 report is useful evidence about a vendor's security and operational controls. It does not tell you what the AI is designed to do, how it was tested for accuracy or unfair discrimination, how outputs can be explained, how model changes reach you, or what data rights the vendor holds. Treat the SOC 2 report as one input to the security part of the review. Your security team or provider should review it, including any exceptions. The AI-specific questions still need their own answers and evidence.
Raise these terms with your counsel: data use and training restrictions; data retention, return and deletion; notice before material model, data or feature changes; audit rights or access to qualified audit reports; cooperation with regulatory inquiries and with your carrier's oversight; incident notice and investigation support; subcontractor and model-provider disclosure; records supplied on request; exit terms that keep past decisions explainable. The NAIC Model Bulletin's third-party guidance specifically mentions audit rights and regulator cooperation "where appropriate and available." Your carrier may expect the same rights from you, so check that your vendor terms let you meet your own commitments. AIGA identifies the contract questions during the review. Counsel drafts and negotiates the terms.
A gap does not have to stop the deal, but it needs a recorded decision. There are four typical outcomes: Approve: the material questions are answered. Approve with conditions: each gap has an owner, a due date and an interim control, such as more human review. Remediate before signature: specific fixes or contract terms come first. Reject: an unresolved gap sits too close to underwriting, pricing or claims outcomes. The decision record captures the outcome, the conditions, who made the decision and when the vendor will be reassessed.
AIGA's two co-founding principals scope and deliver every engagement themselves.
See the Team Bios.
We don't sell or require any AI or compliance software. Every decision and every artifact stays with you.
We work alongside your existing providers. We lead the AI governance workstream so your other partners can focus on what they do best.
Tell us which vendor or use case you are weighing, what is driving the timing (a signature, a renewal, a carrier request or a board meeting), and who needs to see the result. We will come back with a fixed-scope MGA AI Vendor & Procurement Risk Review Sprint proposal, with named deliverables, assumptions and exclusions.
Copyright © 2026 AI Governance Advisors (AIGA) | www.aigovadvisors.ai - All Rights Reserved.
AI Governance for Insurance
MGAs & MGUs