The Review Checklist: Nothing Leaves the Firm Without a Named Reader

Why the rule has to be a name and not a role

Every bank already has a review rule. It says the partner reviews the CIM. The problem is that a role can be assumed and a name cannot. When the rule says the partner reviews, and the partner is traveling, the associate assumes the partner will review it on the plane, the partner assumes the associate would not have sent it without a look, and the CIM goes out with nobody having read page fourteen.

A name closes that gap. If the pipeline entry for the CIM says reader: blank, it does not go. If it says a name, the firm has recorded who was assigned responsibility for the review, and if page fourteen is wrong, the firm knows whose conversation to have. The name is an accountability control, not proof that the artifact is error-free. NIST describes documentation as a way to support transparency and accountability in AI risk management, while FINRA says firms using generative AI should address governance, privacy, integrity, reliability, and accuracy in their policies and procedures. [1] [2]

AI makes this more urgent, not because generated drafts are worse than human drafts, but because there are more of them and they arrive faster. An associate who used to produce one CIM a month and one buyer list a quarter now produces a draft of each in a week. The review step did not get faster. The rule keeps the two in step.

The artifact list: what counts as leaving the firm

The checklist starts with a list of every artifact that reaches an owner, a buyer, or a buyer's advisor. Most banks list five and forget three.

The five everyone lists: the teaser, the CIM, the buyer list the owner approves, the NDA, and the IOI comparison that goes to the owner. The three most banks forget: the client questionnaire as sent, because the owner sees it and an error there propagates everywhere; the weekly owner update, because it is client-facing and is now generated from the pipeline; and the data room index, because buyers read it and it tells them what the firm thinks is important.

Internal artifacts, the normalized intake, the valuation trace, the sorted CIM draft, do not need a named reader before they are used, but they need one before they are relied on. The valuation that supports the teaser's range has a reader. The intake that feeds the valuation has a reader. The chain runs back to the owner's upload.

Write the list down with the reader for each. On a typical mandate, the analyst reads the intake, the associate reads the valuation trace and the buyer list before markup, the senior banker reads the marked-up buyer list, and the managing director reads everything client-facing. That is four names and eight artifacts.

The thirty-second checklist per artifact

For each artifact, the reader answers four questions and then writes their name.

Does every number trace. The reader picks two figures, not the obvious ones, and follows them back to the normalized financials or the data room. If either breaks, the artifact goes back. This takes a minute on a traceable valuation and an hour on an untraceable one, which is itself information about the tool.

Is there anything generic. The reader scans for a paragraph that could describe any company in the vertical. One is tolerable in a teaser. One in the CIM's growth section is not.

Is there anything the owner would not say. The reader looks for a claim about the business that the owner, reading it, would dispute or be surprised by. This is the question that catches the most, because generated drafts are assembled from the data and the data does not know what the owner is sensitive about.

Who else has to see this before it goes. Usually nobody. Sometimes the owner, for a buyer list. Sometimes counsel, for an NDA with a changed clause. The reader writes that name too, and the artifact waits for the second name.

Then the reader's name goes next to the artifact in the pipeline, and the associate sends it.

What happens when something gets through

Something will. A CIM will go out with a customer named that should have been anonymized, or a buyer list will include a competitor the owner asked to exclude. The named-reader rule does not prevent every error. It changes what happens next.

Without a name, the error produces a meeting about process. With a name, it produces a conversation between the managing director and the reader about what they looked for and what they missed, and a line in the SOP that adds that check to the thirty seconds. The customer anonymization check becomes question five. The excluded-buyer check becomes a step before the senior banker's markup.

The checklist grows by one line per error, which is slow, and it grows from what actually went wrong, which is the only growth worth having. A checklist written in advance by a committee has twenty questions nobody asks. A checklist grown from a year of mandates has seven questions everybody asks.

Keeping the rule when the firm is busy

The rule is easiest to keep in month one and hardest to keep in month six, when four mandates are in outreach at once and a buyer wants the CIM by end of day. The associate has the draft, the partner is on a call, and the temptation is to send it and write the partner's name afterward.

Two things prevent that. The first is that the pipeline makes the blank visible. If the reader field is empty, it is empty on the Monday screen, and the sponsor sees it. The second is that the associate was told, in the first training session, that writing someone else's name without their reading is the one thing that ends their time at the firm. Say it once, plainly. It only needs to be said once.

The rule bends in one way only. If the managing director cannot read the CIM by end of day, another partner can, and their name goes on it. The rule is a named reader, not a specific one.

This week

Pull up the last three artifacts that left the firm on any live mandate. For each, ask who read it before it went. If the answer is a role rather than a person, or a person who is not sure, write the rule into the pipeline on Monday and start applying the four questions.

How Acquivo Pro fits

Acquivo Pro gives the named-reader rule a place to live. Every client's artifacts move through a single pipeline, from the questionnaire the owner answers in the branded portal through valuation, teaser, CIM, prospect list, NDA, IOI, and LOI, so each one has a visible status that a reader's name can sit beside. The valuation engine is deterministic and traces every figure to source, which turns the first checklist question into a one-minute click-through instead of a rebuild. Client data is isolated and not used to train models for other customers. To see it on a live mandate, request a demo through the Contact page.