On 22 September, Citi published its 2026 Global Family Office Report. This year, the report pays particular attention to how family offices are using artificial intelligence, providing a much clearer picture of where the market stands today.

The numbers are telling. 40% of family offices describe themselves as being at an early stage of AI adoption, 30% at a moderate stage, 10% at an advanced stage, and a further 2% describe their use of AI as transformational. In other words, the question of whether family offices need AI is gradually giving way to a more practical one: how should they actually use it? Adoption still varies considerably, with larger family offices noticeably ahead of smaller ones.

At CGL, we looked at Citi’s findings from this perspective: where is AI already delivering practical value to family offices, and where is that value still emerging?

Productivity is delivering the clearest results

While much of the earlier attention around AI focused on its potential role in investment management, the clearest measurable benefits today are operational. Family offices are using AI for meeting notes and summaries, workflow automation, document processing, knowledge management and research assistance. According to Citi, 35% already report reductions in manual work or headcount. For smaller family offices, this is particularly significant.

A small family office faces many of the same demands as a much larger organization: investments, reporting, multiple external providers, documentation, compliance, and often structures spanning several jurisdictions. But it has far fewer people to manage them. AI therefore offers a relatively straightforward opportunity: a small team can do more without increasing headcount at the same pace.

Investment adoption is developing more gradually. Citi identifies applications across market and macro analysis, due diligence, portfolio and risk monitoring, deal sourcing and opportunity screening, as well as automated reporting and performance analytics. But adoption remains at an earlier stage here, particularly as AI moves closer to actual investment decisions and alpha generation.

This does not necessarily indicate a problem with AI adoption. Rather, it may reflect where family offices are currently comfortable using the technology. AI can analyse information, identify patterns, assist with due diligence and support investment professionals in making decisions. Allowing it to make decisions over family capital, however, is a very different proposition. Part of this may simply be habit. Part may reflect continuing concerns around AI explainability and accountability. And part may be inherent in the family office model itself: families create these structures precisely because they want to retain control over their wealth.

At the same time, another trend is emerging

While investment applications are developing gradually, Citi’s findings reveal a particularly strong cluster around information management and reporting. AI is already being used for document retrieval and processing, internal knowledge management, research assistance, automated reporting and performance analytics. And the strongest measurable benefit in Citi’s AI findings is directly related to information: 67% of respondents using AI report faster information processing. By comparison, 19% report improved reporting accuracy. This also reflects what we at CGL increasingly hear from professionals working with family offices. There is a practical reason for this.

Family office reporting is often complex, time-consuming and expensive. A family may have companies, investment portfolios, bank accounts, funds, real estate and other assets spread across multiple jurisdictions. Information comes from different banks, asset managers, accountants and advisers, often in different formats and subject to different local requirements. The family office somehow has to bring all of this together.

At the same time, its own managers cannot realistically be experts in the accounting rules and practices of every jurisdiction in which the family operates. This makes it difficult to review the work of multiple external providers and creates an obvious information asymmetry between the family office and the specialists it pays. AI offers a potential way to reduce that asymmetry by helping to collect, structure, compare and analyse information, while giving family office professionals a greater ability to ask the right questions and verify what they receive from external providers. This is a far less controversial role for AI than managing family investments — and one where the economic value may be much easier to demonstrate.