Wilmington
The Caremark duty now reaches AI.
Delaware doctrine holds directors personally liable for failing to maintain a functioning system for monitoring known risks. AI is now a known risk. “We hadn’t gotten to it yet” is not a system.
AI oversight is now a fiduciary matter — Delaware doctrine, SEC disclosure, D&O renewals. We put a working answer in front of your board in a single agenda slot: what management is actually running, what it exposes you to, and the questions to ask before someone else asks you.
Prefer to go direct? The principal reads his own mail: paul.washburn@sovereign-action.com
Board of Directors — regular meeting
Agenda · third quarter
Item 7 — deferred Q4, deferred Q1, deferred Q2.
66%
of directors report limited or no knowledge or experience of AI
NACD, 2026
9%
of companies have a formal, board-adopted AI policy
ISS STOXX, 2026
6%
of boards receive AI reporting metrics from management
NACD Board Practices Survey, 2025
+10.9
points of return-on-equity advantage for boards with AI-savvy directors
MIT CISR
Meanwhile 88% of organizations already run AI in at least one business function. The gap between what companies are doing and what boards can see is the widest it has been since cybersecurity in 2016 — and it is closing the same way: through liability.
Wilmington
Delaware doctrine holds directors personally liable for failing to maintain a functioning system for monitoring known risks. AI is now a known risk. “We hadn’t gotten to it yet” is not a system.
Washington
The SEC has brought enforcement against AI-washing, and its Investor Advisory Committee has recommended that companies disclose the board’s AI oversight mechanisms. The reporting season that pattern produces is predictable — cybersecurity ran the same script.
Your insurer
Underwriters have begun adding AI governance to renewal questionnaires. The renewal where the board has no documented answer is the expensive one.
None of this requires your board to become technologists. It requires a documented system of oversight — questions asked, reports received, decisions minuted.
Prices are published because your time is worth more than a discovery process. Each engagement stands alone; each one also pays for the next decision to be an easy one.
$12,500 · fixed fee
One agenda slot · ninety minutes · your boardroom or remote
The full board, one session, no theater. The liability landscape as it stands, a live demonstration of what AI agents actually do inside a company like yours, and the questions your board should be putting to management — before someone else puts them to you.
Credited in full against an assessment commissioned within ninety days.
The board keeps
From $45,000
Four to six weeks · board-grade written report
A structured reading of the company’s actual AI estate: what management has deployed, what’s running in the shadows, what it exposes the company to, and where the unclaimed value sits. Findings are demonstrated live on the company’s own systems — evidenced, not asserted in a maturity matrix.
One assessment at a time. Scope and fee fixed before work begins.
The board keeps
From $6,000 / month
Quarterly attendance · on call between meetings
A standing seat at the table without a search process. Written brief before each meeting, attendance at the AI agenda item, and a direct line between meetings — for the vendor pitch that sounds too good, the incident that needs a straight read, the regulator’s letter.
Four seats maximum, held concurrently. Quarterly, in advance.
The board keeps
Twelve questions that separate a board that governs AI from one that receives presentations about it.
Resolution and charter text your corporate secretary can lift directly — the documented system Delaware looks for.
What management reports on AI, to whom, how often. The 6% of boards that have this are the ones sleeping well.
Not a framework diagram — a working view of the company’s AI activity, built on infrastructure we run in production ourselves.
Paul M. Washburn operates Sovereign Action’s own production AI infrastructure — sixty-five shipped systems, including the governance layer itself: compliance scanning, guardrails, audit trails, agent scorecards, drift monitoring. When your board is briefed on agent risk, the risk is demonstrated live on systems we run — not illustrated on a two-by-two.
That is the entire pitch. Most board advisors have read about the machine. Some have audited the machine. We build it, run it, and get paged when it misbehaves — which is precisely the vantage point a board needs from the person telling it what to worry about.
Not a sales rep — there isn’t one. If the board doesn’t need this yet, we say so on the call.
Standard confidentiality both directions. We disclose our commercial interests in writing before any engagement.
Pre-read a week ahead. Ninety minutes in the room. Item seven, finally handled.
No junior staff · no discovery workshops · references on request · discretion by default
Twenty minutes with the principal is enough to know whether your board needs a briefing this quarter or a note in the minutes that it considered the question. Either way, you leave with an answer — and it stays off your successor’s agenda.