Executive Leadership
What a Technology Board Advisor Actually Owes a CEO
Most advisory boards are a row of headshots on an investor deck. Here is what the job looks like when someone takes it seriously, from both sides of the table.
A founder once sent me a spreadsheet of her eleven advisors. Next to each name was the equity grant, the date of the last conversation, and a column she had labeled “changed a decision?” Nine rows said no.
She wasn’t angry about it. She was curious, in the way good operators get curious about any line item that keeps renewing without producing anything, and she wanted to know whether the problem was the advisors, the structure, or her. My honest answer was mostly the structure, a little bit her, and in two cases the advisors, who had accepted equity for a title and then treated the title as the whole job.
I have sat in fiduciary board seats and in advisory seats, and I have been a CEO on the receiving end of both. The advisory role is the one almost nobody defines. That vacuum is why so many of them fail quietly.
An advisor has no vote, and that is the point
A fiduciary director owes duties to the company and its shareholders. They approve budgets, hire and fire the CEO, sign off on financings, and can be sued for getting it badly wrong. Their questions, however friendly, carry the weight of someone who can eventually say no.
An advisor has none of that power. Good.
Because an advisor cannot block anything, the CEO can tell them things that would be premature or politically expensive to say in a boardroom: that the CTO is underwater, that the AI roadmap is mostly a slide, that a key customer has been hinting about leaving since spring, that the Series B story depends on a metric everyone is quietly nervous about. The advisor’s entire value sits in that gap between what the CEO knows and what the CEO is ready to say formally. An advisor who behaves like a shadow director, asking for board decks and grading performance, closes the gap and destroys the reason to have them.
What the technology seat is for in 2026
Five years ago a technology advisor mostly helped a non-technical CEO interview engineering leaders and sanity-check cloud bills. That work still exists. It is no longer the main event.
The main event is AI operating decisions, and CEOs are making them faster than their boards can learn the vocabulary. Should the support team be rebuilt around agents or augmented by them? Who owns an agent fleet when it spans marketing and operations? What does a reasonable monthly model spend look like against the revenue it touches, and when is the vendor telling the truth about accuracy? A board will eventually ask versions of the oversight questions directors should be asking. The advisor’s job is to get the CEO to credible answers before that meeting, privately, with room to be wrong first.
So the advisor has to have actually run the thing. Operated it, broken it at two in the morning, rebuilt it, and paid the invoice when the model bill came in at triple the forecast. I would weigh one advisor who has operated a production agent fleet over three who have keynoted about one.
Scope the engagement like a contract, because it is one
Most advisory relationships begin with a warm phone call and a standard grant, and they never get more specific than that. The ones that work get written down. When I take an advisory seat now, the one-page agreement answers these:
- The problem. One named question the CEO needs help with over the next two quarters, written in a sentence. “Decide whether to consolidate our three AI vendors into one, and build the internal team to run it” is a problem. “Help with AI strategy” is a horoscope.
- Cadence. A fixed monthly hour plus open access by text for anything urgent. The standing hour matters less than the text thread, which is where most of the real advising happens.
- Access. Whether the advisor can talk directly to the CTO, the head of marketing, or the team running the agents, and whether the CEO wants to hear about those conversations.
- The exit. A review date, usually at twelve months, where either side can end it without a conversation about feelings.
Equity follows scope, not reputation. The common range for an early-stage advisor is still roughly a quarter to one percent over two years with monthly vesting and a short cliff, and the upper end should be reserved for advisors doing close to fractional-executive work. If the engagement is mostly introductions, pay for introductions. If it is sitting with the team while they redesign how work gets routed to agents, that is a different amount of value, and the grant should say so.
How to tell whether yours is working
Borrow the founder’s spreadsheet column. Every quarter, list the decisions you made that mattered and mark which ones an advisor changed or talked you out of. A single entry per advisor per quarter is a healthy relationship.
Zero entries for two quarters in a row means the scope is wrong or the person is. Either way, have the conversation. The kindest thing a CEO can do for a coasting advisor is let them go early, while the relationship is still intact and before the vesting schedule turns the whole thing into an awkward annuity.
A few warning signs show up long before the spreadsheet does. The advisor answers questions you did not ask. Every recommendation happens to route through a portfolio company or a vendor where they used to work. They want the board deck. And the subtle one, which took me years to notice: they are always agreeable, which in an advisor usually means they stopped paying close attention some time ago.
What I owe the CEOs I advise
I owe them the uncomfortable sentence first, before the context that softens it. I owe them a straight answer about whether I have seen the problem before or am reasoning from adjacent experience, because those deserve very different levels of trust. When a question lands outside what I know, I owe them a name instead of an opinion.
And I owe them silence in the boardroom about anything they told me in confidence, even when I also know a director, even when I think the director should hear it. If the CEO is hiding something material from the board, the right move is to tell the CEO that directly and, if needed, step away from the role. Leaking it would be a betrayal of the one thing an advisor has to offer.
Most weeks the job is small. A text about whether a vendor’s accuracy claim sounds plausible. A fifteen-minute call before a hard conversation with a CTO. A reread of the governance framework the CEO is about to take to the board, with three lines marked as things a skeptical director will pull on.
Small, and worth a great deal, if the advisor has done the work.
This article is part of the Executive Leadership cluster, focused on board governance and the operating discipline required to run AI systems responsibly at the executive level.