The question I am asked most often by board chairs of mid-sized companies considering their AI governance gap is a version of this: “Do we need to appoint a full NED with technology expertise, or can we use a fractional advisor?”
It is a reasonable question. The cost difference is material — a full non-executive director in the UK typically costs GBP 30,000-70,000 per year in fees for a mid-sized company, plus the time and process involved in a formal appointment. A fractional technology advisor is typically GBP 5,000-20,000 per year depending on scope and engagement frequency.
But the question, framed as a cost comparison, misses the governance dimension that makes it genuinely important. This is not a budget decision. It is an accountability decision.
What a NED provides that an advisor cannot
A non-executive director has a fiduciary obligation to the company. They attend every board meeting as a member of the governance body. They vote on board decisions. They are legally accountable, alongside their fellow directors, for the governance quality of the board’s decisions. If an AI governance failure reaches a regulator, the NED is in the room where the decisions were made, with the accountability that goes with board membership.
An advisor provides expertise. They can attend meetings by invitation, provide written assessments, review proposals, and recommend courses of action. They do not vote. They do not carry fiduciary accountability. When the regulator asks who was responsible for the board’s AI governance decisions, the advisor’s name is not on the board composition record.
For companies where AI governance is primarily a knowledge gap — the board needs someone who understands AI well enough to ask the right questions and evaluate management proposals — a fractional advisor is often the right first step. It is faster to engage, more flexible to adjust, and proportionate to a board that is building AI governance capability rather than operating an established one.
For companies where AI governance is a liability question — where the board has approved AI deployments that fall under EU AI Act Annex III, where NIS2 personal liability is in force, or where the company is in a regulated sector with material AI governance obligations — the advisor’s non-fiduciary status is a limitation. The board needs a member who carries the governance accountability, not a consultant who carries the advice relationship.
The practical difference in the boardroom
There is a practical difference between how an advisor and a NED function in board AI governance that is worth understanding before making the choice.
An advisor’s input is advisory. The board can take it or leave it. If the advisor recommends against approving a deployment and the board approves it anyway, the advisor’s view is noted and the board’s decision stands. The advisor has no mechanism to enforce the governance standard they have been engaged to provide.
A NED who votes against approval is a different matter entirely. A dissenting NED vote is a governance event: it is recorded in the board minutes, it creates accountability on both sides, and it is visible in the board’s governance record. A regulator or a shareholder reviewing the board’s AI governance can see whether the AI expert director dissented on a decision that later became problematic.
This accountability function is what the fiduciary relationship provides. It is not primarily about whether the NED is more or less knowledgeable than the advisor. It is about whether the AI governance expertise is embedded in the board’s decision-making accountability or whether it is available as an input to the board’s decisions.
The hybrid approach most mid-sized companies should use
The choice is not binary. The most effective governance structure for many mid-sized companies in the AI governance transition period is a combination:
An external fractional advisor with deep AI and technology governance expertise, engaged for 12-18 months to build the board’s collective AI literacy, establish the governance framework, and assess AI readiness. This is a capability-building engagement, not a permanent governance structure.
Followed by a formal appointment of a technology NED, selected from a candidate pool that has been identified as part of the capability-building period, to take on the fiduciary governance function once the board’s governance structures are established.
The sequencing matters. Appointing a technology NED before the board has established its AI governance framework asks the NED to build the framework as well as provide the oversight — which is a significantly larger scope than providing oversight of an existing framework. The fractional advisor builds the framework. The NED operates it.
The cost calculation done correctly
The full cost of a technology NED for a mid-sized company is not just the annual fee. It includes the nomination and appointment process (typically 3-6 months), the induction period, and the board time involved in integrating a new member into the governance structure. For a board that does not yet have an AI governance framework, the NED appointment also includes the time for the NED to build that framework.
The full cost of a fractional advisor is not just the annual fee. It includes the limitation: if the board needs a governance voice with fiduciary accountability, the advisor cannot provide it.
Done correctly, the cost comparison is:
- Fractional advisor + 12-18 months to build framework + NED appointment = the right structure for a board starting from a governance gap
- Direct NED appointment without framework = higher annual cost, slower time to governance capability, risk that the NED is building framework instead of providing oversight
- Fractional advisor indefinitely without NED = lower cost, no fiduciary governance accountability, growing regulatory liability as the AI Act enforcement window approaches
The cheapest option over a three-year horizon is the first one. The most common option is the third — and it is the most expensive over the same horizon when the regulatory liability is correctly priced.
The Board AI Governance Framework provides the governance structure that a board needs to establish before a technology NED appointment — or alongside a fractional advisor engagement. It is the foundation that makes either relationship more effective.
For boards seeking independent advisory support on AI governance and technology board composition, contact Steven directly.
