Executive Advisory: an independent counterpart for major technology decisions
We sit beside the CEO and the board while an AI or technology decision is made, test what it rests on, and leave a record of the reasoning, the owner and what would reopen it.
In brief
Executive advisory gives a CEO or board member an independent senior counterpart for AI and technology decisions. We read the papers before the meeting, test what the decision rests on, say plainly where we disagree, and write down the choice, the owner and the evidence that would reopen it. The decision stays with you.
On this page
What has already happened to a decision by the time it reaches the top team?
By the time a paper reaches the CEO or the board, it has been shaped. The sponsor chose the evidence, the vendor supplied the benchmarks, the delivery partner estimated the work, and finance fitted the numbers to the case. None of this is improper. Each person describes the decision from where they stand.
The effect is that the room receives a finished argument, and the questions that would test it were never put by anyone placed to ask them. Months later the decision returns: results disappoint, a new executive arrives, a director remembers a concern. The reasoning cannot be found, or it can be found and nobody can say what would have changed it. The debate starts again on opinion.
Who sees which part of a decision?
Every voice in the room is useful and none of them is placed to see the whole.
| Voice in the room | What it sees best | What it cannot see from where it sits |
|---|---|---|
| The sponsoring executive | The need, and the cost of doing nothing | The alternatives that would change the case |
| The vendor | What the product does at its best | How your systems and your people will receive it |
| The delivery partner | The effort, as its own method would perform it | Whether the work should be done at all |
| The finance partner | Whether the numbers hold together | Whether the assumptions under the numbers are right |
| A board member | The pattern from other companies | The detail of your systems and constraints |
| An independent counterpart | The decision as a whole, with no stake in the answer | Anything the papers and the owner do not show us |
A decision, from first paper to written record
- 01
Read the paper before the meeting
The sponsor sends the draft. We return a brief setting out the evidence relied on, the evidence missing, the assumptions that carry the argument and the questions the room should ask.
- 02
Speak with the owner alone first
A private conversation with the executive who will answer for the decision, where assumptions can be questioned without an audience. This is where we say plainly if we think an option is wrong.
- 03
Be in the room, or behind the chair
When asked, we attend to present the brief or to answer questions on the paper. We do not vote, and we do not speak for management.
- 04
Write the decision down
Within days there is a record: the choice, the owner, the reasoning, the options set aside and why, the assumptions it depends on, and the conditions that would reopen it.
- 05
Return at the checkpoint
On the date the record set, we check whether the assumptions still hold. If they do, the decision stands without another debate. If not, the record says what changes.
A decision nobody can reopen on evidence will be reopened on opinion.
A pattern we are called into
A hospitality group asked to consolidate booking, guest profiles and loyalty
The chief technology officer sponsors a single platform for booking, guest profiles and loyalty. The vendor has given a convincing demonstration and the delivery partner has estimated the work.
The CEO senses the case is stronger than the group's own history with large migrations would suggest, but cannot say why.
- Read the paper. The estimate assumes every property will adopt one booking flow. The paper does not say who has authority to require it.
- Talk to the owner. The chief technology officer agrees the adoption assumption is the weakest point, and had not yet had occasion to say so.
- Brief the board. We set the decision out as a staged commitment: what must be true after the first stage for the second to be approved.
- Write it down. The record holds the owner, the staged commitment, the adoption evidence that would release the next stage, and the checkpoint.
Where it endsThe board approves the first stage only. The next conversation is about the adoption evidence, not about whether anyone believes the sponsor. In another version of the same pattern, the same reading ends in advice not to proceed, and the record says so.
When do we advise against proceeding?
Sometimes the right result of a review is a recommendation not to proceed, or to proceed on a smaller commitment. We say that when the reasoning cannot be made to hold: an assumption that carries the case has no evidence behind it and cannot cheaply get any; the commitment cannot be undone and the owner cannot say what would make it a mistake; or the decision depends on a capability the business has no plan to build.
We say it in the pre-read, to the owner first, and the record carries it whatever the board decides.
Is an independent counterpart what this decision needs?
It fits when
- One large or hard-to-reverse decision is coming and the evidence has come mainly from the people who want it
- A new CEO, an acquisition or a change on the board has reopened settled questions
- The board has asked for an independent view before it approves
- A decision keeps returning and nobody can say why it was made
Look elsewhere when
- You want the decision confirmed and the debate ended, because a test that cannot change the answer is not advice
- You need someone to run the work, which is delivery in its own scope
- The question is legal, tax or audit, which belongs with the specialist adviser
- The executives who decide will not give direct access, since the relationship depends on it
The cost is exposure. A sponsor's argument is tested before colleagues see it, and written records make a decision harder to rewrite later. Leaders who value that tend to value it more the second time.
What does a standing relationship ask of you that a single decision does not?
A single decision needs the papers, a private conversation, an appearance if the board wants one, and a little time afterward for the record.
A standing relationship keeps a live decision list with owners and dates, a regular session with the executives who own those decisions, and a review of board and committee papers before they are presented. It asks more of the CEO's attention and returns a calmer top team: fewer decisions arrive unprepared, and the old ones stop coming back.
What do you hold afterward?
- A standing decision list with an owner and a date against each item
- Pre-read briefs: the evidence, what is missing and the questions for the room
- Written decision records with reasoning, assumptions and reopening conditions
- Comments on board and committee papers before they are presented
- A statement of any interest we have in an outcome, made before we advise on it
What do we need, and what do we leave to you?
We need direct access to the executives who decide, and the papers a decision requires, held under the NDA that covers every engagement. We do not manage the business, commit it to anything, vote on a board or speak for management.
A decision we advised against may still be taken. If so, the record holds our view beside yours and the conditions under which each of us would change our mind. We do not guarantee that a decision will succeed. We agree how it will be reviewed.
Where can you frame a decision before we meet?
The investment decision brief sets out the questions we put to any AI commitment, and the sponsor can complete it alone. The Field Report pieces on decisions that return and on actions that go unowned describe the failures this work is built to stop.
Questions a CEO or director asks before engaging us
How do you keep what we discuss confidential?
Every engagement is under NDA, and we ask only for the papers a decision needs. Notes are held under the same controls as all client material, which the Trust Center describes, and nothing from one client's decision informs another's.
What if we disagree with your advice?
Then the decision is yours and the record says so. We write down our view, your reasoning and the conditions under which each side would change its mind, so the disagreement becomes information the next review can use.
Can you attend board meetings?
Yes, when asked, to present or to answer questions on a paper we have reviewed. We do not vote or speak for management, and the scope states which meetings we attend and what we prepare beforehand.
How is this different from a non-executive director or a board adviser?
A director carries legal duties, a vote and a view across the whole company. We carry none of those and do a narrower job: reading a specific decision closely, testing it against your systems and delivery record, and writing it down. Some boards use both.
Related
Who tests the decision before the money is committed?
Describe what you are weighing in a few lines. A senior practitioner replies with a suggested first conversation.