Over the past two years, "Head of AI" has become one of the most requested leadership titles across Hong Kong. Banks, insurers, conglomerates, utilities and family offices all want one. Many of these searches run for months without a hire. In most cases, the problem is not a lack of talent. It is the brief.

One title, three different jobs

When we unpack a typical Head of AI brief, it usually describes three distinct roles:

  • The strategist. Someone who can sit with the executive committee, identify where AI genuinely changes the business, build the roadmap, and handle governance, risk and regulatory questions.
  • The builder. A technical leader who can stand up data and machine learning platforms, hire engineers, and get models from prototype into production safely.
  • The adopter. A product and change leader who makes sure the tools are actually used by the business, measures value, and drives the process change that makes the investment pay off.

Each of these is a real job. Each has its own talent pool, its own career path and its own compensation level. Very few people do all three well, and those who do are rarely open to a move.

Why that stalls the search

When a brief asks for all three, a predictable pattern follows. Strong strategists are rejected for not being hands-on enough. Strong engineers are rejected for not being commercial enough. Strong product leaders are rejected for not having a deep technical background. The shortlist keeps resetting, and the role stays open.

Compensation adds to the problem. A budget set for one profile rarely matches the market rate for another. The engineering leader with production experience at scale and the transformation leader who can work with a board are priced differently, and candidates know it.

Hong Kong adds its own constraints. The pool of leaders who have taken AI into production in a regulated environment here is still relatively small. Many of the strongest candidates are in Singapore, mainland China or further afield, which brings relocation, family and visa questions into the search early.

What works better

  1. Decide which job comes first. If the organisation has no data foundations, the first hire is probably the builder. If pilots exist but nobody uses them, it is probably the adopter. If the board is still asking what AI means for the business, it is the strategist.
  2. Price the role you actually need. Benchmark against the profile you are hiring, not against the title.
  3. Plan the team, not just the leader. The first hire should be someone who can recruit or partner for the other two capabilities over the following eighteen months.
  4. Be honest about the starting point. Good candidates will ask detailed questions about data quality, budget and executive sponsorship. Clear answers help close them. Vague answers lose them.

The real question

The question is rarely "where do we find a Head of AI?". It is "what do we need AI to do for us in the next two years, and who is the right first hire to start that?". Answer that well, and the search becomes much shorter.