1. Why This Document Exists
New product development at an insurance carrier spends capital and scarce specialist capacity long before it produces a single policy. The pricing actuary modeling a rider is not available for in-force experience work while she does it. That is the real cost of an NPD program, and it is why this one is governed by a board rather than by a sponsor.
An org chart shows the thirteen teams. What it cannot show is the thing that actually determines how decisions get made here: who is in the room, who has a vote, and who deliberately does not.
⚠ C. Tyrrell chairs the Gate Review Board and does not vote (decision record D-11). He is seated from the PMO and owns gate integrity and evidence quality — which is precisely why he must not rule on the evidence he assembled. ***The person who controls what the board sees cannot also be the person who decides what it means.***
⚠ D. Pemberton, Chief Risk Officer, attends every gate as a non-voting observer — but with an unrestricted right to record a dissent in the gate minute, escalated to the Executive Committee within five business days. ***A recorded dissent does not block the decision. It makes the decision permanently attributable, which is a different and often stronger control.***
One has no vote because he is too close to the evidence. The other has no vote because he is deliberately outside the decision and must still be able to object on the record.
2. The Gate Review Board
Five voting seats, two non-voting. ⚠ Quorum must include the Chief Actuary and the General Counsel & Chief Compliance Officer — a gate cannot be decided without the two functions most able to stop it.
| Seat | Holder | Owns at the gate | Vote |
|---|---|---|---|
| Executive Sponsor | G. Marchetti, Chief Product Officer | Product viability, portfolio fit | Voting |
| Chief Actuary | N. Adeyemi | Pricing, reserve adequacy, assumption integrity | Voting — quorum-critical |
| General Counsel & CCO | B. Lindqvist | Filing, contract forms, suitability, market conduct | Voting — quorum-critical |
| Head of Distribution | R. Castellanos | Volume assumptions, channel capacity | Voting |
| Chief Financial Officer | J. Whitmore | Capital, hurdle rate, funding release | Voting |
| Chief Risk Officer | D. Pemberton | Enterprise risk | No vote — dissent right |
| Chair | C. Tyrrell, NPD Program Manager | Gate integrity, evidence quality | No vote |
⚠ Seats are held by role, not by individual. A member may send a delegate, but only one holding written authority. That rule exists so a gate is never quietly decided by whoever happened to be free that morning.
3. What the Chair Actually Owns
Removing the vote does not reduce the role; it relocates it. The chair owns everything about the gate except the decision, and that is a substantial and specific list:
- The decision package — that entry criteria are genuinely met before a gate convenes, rather than convening and discovering they are not.
- Evidence quality — that what the board is shown is complete, current, and not selectively assembled to produce a preferred outcome.
- Gate conditions — issuing them, tracking them, verifying closure. A condition that is issued and never verified is worse than no condition, because the record says it was handled.
- The minute — including any CRO dissent, recorded whether or not it is welcome.
4. The Delivery Organization
Thirteen teams, each with a lead, and each owned at board level by a functional executive. ⚠ That second column matters more than it looks: every team has a route to the board that does not run through the program, which is what stops the Program Manager becoming the sole channel between the work and the people deciding its fate.
| Team | Lead | Board owner | Owns |
|---|---|---|---|
| Actuarial — Pricing & Product Development | S. Ravichandran | N. Adeyemi | Rider pricing, assumption setting, profitability modeling |
| Actuarial — Valuation & Financial Reporting | T. Brennan | N. Adeyemi | Reserve methodology and reporting treatment |
| Legal & Compliance | A. Nkemelu | B. Lindqvist | Suitability, market conduct, contract language |
| Regulatory Filing & Contract Forms | P. Hollingsworth | B. Lindqvist | State filings and the filing calendar |
| IT — Annuity Administration Platform | V. Sandoval | G. Marchetti | The largest single build (10 people) |
| IT — Illustration & Quoting | E. Kowalczyk | G. Marchetti | Illustration engine and quoting accuracy |
| Product Management & Design | D. Falkner | G. Marchetti | Product construct and feature set |
| Operations — New Business & Policyholder Services | L. Marchand | G. Marchetti | Issue, service, and the operational readiness case |
| Investments — ALM & Hedging | M. Delacroix | J. Whitmore | Hedging strategy and asset-liability fit |
| Data, Reporting & Analytics | F. Osunde | J. Whitmore | Reporting infrastructure |
| Distribution & Wholesaling | H. Kirkpatrick | R. Castellanos | Channel readiness and wholesaler enablement |
| Product Development PMO | K. Iyer | C. Tyrrell (Chair) | Plan, evidence assembly, gate logistics |
5. The Two Refusals — Recycle and Cancel
Most governance models have one way to say no. This one has two, and the distinction is the most useful idea in the suite. ⚠ A recycle returns the package without deciding the product. It exists so the Board is never forced into a false choice between approving something unproven and canceling something that may be sound — the two failure modes a single-outcome gate produces.
| Recycle | Cancel | |
|---|---|---|
| What it decides | Nothing about the product. Only that the evidence is not yet decidable. | The program ends. |
| Who tables it | The Chair — who has no vote | Any voting member; available at every gate |
| What the Board must record | The specific deficiencies and the date the package returns | Final accounting within 20 business days to the CFO and Executive Committee |
| Money | Stage tranche not released; the program draws only on current funding. Any overrun from the loop is a contingency draw requiring Board authorization. | Funding stops immediately; unreleased tranches canceled; unspent funds returned to the capital pool |
| People | Team continues | Released on a stated date, reassignment through the functional line |
| Vendors | Unaffected | Committed obligations are settled, not abandoned — termination-for-convenience exercised under the Vendor Management Plan |
⚠⚠ Note who holds which lever. The Chair cannot vote on a gate outcome, but he is the one who tables a recycle recommendation — and cancellation is available between gates on a Chair-convened extraordinary session. ***The person with no vote is the person who can call the meeting at which the program dies.*** That is not a contradiction. Facilitation authority and decision authority are different things, and this design separates them deliberately: he can force the question in front of the people entitled to answer it, and he cannot answer it himself.
6. What the Program Cannot Decide
Pricing and reserve adequacy. The Chief Actuary's judgment on assumption integrity is not a program input to be balanced against launch date. An assumption the actuary will not sign is not a schedule problem.
Filing and suitability. A product that cannot be filed cannot be sold, and no amount of program pressure changes a state's requirements.
Funding release. Money is released one tranche at a time by the CFO at gates. The program does not draw ahead of a gate, which is the mechanism that makes cancellation genuinely possible rather than theoretically possible.
7. How This Compares To the Other Programs Here
| Question | Lighthouse (this suite) | Catalyst (AI) | PM / Agile |
|---|---|---|---|
| Where does the PM sit? | Chairs the board, no vote | Not on any board | Chairs, and decides |
| Who can stop it? | Chief Actuary, GC/CCO, CFO — each alone | Independent Model Validation, unappealably | The sponsor |
| How is refusal recorded? | A CRO dissent in the minute, escalated in 5 days | A failed validation gate | A declined change request |
| What is protected? | Capital, and the policyholder behind the pricing | The member behind the model | The delivery commitment |