Nassau Street Partners: Family offices take centre stage in private markets
There was something revealing about the way artificial intelligence was discussed at the 2026 Nassau Street Partners Family Office Summit in London. The conversation did not begin with grand predictions about machines replacing investors or autonomous systems remaking global finance. It began with practical questions. Where can the technology save time? Where can it identify information that older tools miss? Where does it remain unreliable? And, most importantly, who should exercise judgment when the model reaches the limits of its competence?
Those questions are especially relevant to family offices, which are becoming more important participants in private markets. Once viewed mainly as discreet stewards of a single family’s wealth, many now operate with the breadth of small institutions. They evaluate direct investments, co-invest alongside funds, assess private credit opportunities, monitor operating businesses and build relationships with managers across sectors and geographies. Their advantage is often flexibility. Their challenge is that flexibility creates an enormous volume of information to process.
That tension framed the London discussion. Nassau Street Partners, led at the event by chairman Gary Shields, brought together financial professionals and family-office representatives for a candid exchange with Peter Danenberg, a senior software engineer at Google DeepMind. Danenberg’s remarks were notable for their restraint. Large language models, he said, were not suitable for every task. In high-frequency trading, latency can rule them out. Yet the same systems can prove useful in areas where speed is measured in minutes or hours rather than microseconds.
Sentiment analysis was one example. Danenberg described how an LLM could be directed at an earnings call and used to interpret tone and language with results that may materially improve on older methods. Compliance was another. The ability to read lengthy documents, compare requirements and flag potential issues before a trade may lack the glamour of an automated trading strategy, but it can protect capital and reduce expensive mistakes.
For family offices, this practical emphasis matters. The private markets reward investors who can form a view before an opportunity becomes widely intermediated. That does not necessarily mean moving faster than everybody else. It can mean organizing information more effectively, asking better questions and deciding which risks deserve deeper attention. An attendee described using AI to challenge investment memoranda against investment-committee guidelines. The model was not treated as an oracle. Its value lay in acting as a tireless critic, identifying omissions or weaknesses that a busy team might otherwise overlook.
This is where the growing influence of family offices becomes clear. In public markets, a large amount of information is standardized, continuously priced and distributed almost instantly. Private investments are different. The documents are less uniform, the history may be incomplete, the assumptions can be highly specific and the quality of management often matters as much as the spreadsheet. The investor has to combine financial analysis with context, relationships and judgment.
AI can help structure that work, but it cannot remove the need for experienced decision-makers. Danenberg compared the process to pointing a model at a vast software codebase and asking it to find every bug. The system may generate hundreds of suggestions, most of which are irrelevant or wrong. The useful output appears only when a human can identify the small number worth pursuing. The same principle applies to private-market diligence. A model may produce a long list of concerns, but it takes judgment to distinguish a genuine threat from a theoretical one.
The summit also highlighted a change in how family offices think about internal capability. One participant described an effort to build an AI version of himself that could help him work more quickly and train junior colleagues. The ambition was not to allow the system to advise clients independently. It was to preserve and distribute hard-won knowledge inside the organization. For a family office facing shortages of specialist talent, that is a compelling use case.
Yet the discussion never slipped into easy optimism. Models need to be grounded in the right documents and data. Long files often require preprocessing and careful organization. Important information can become less visible when it sits in the middle of a large body of text. The quality of the final answer depends heavily on the preparation that comes before it. In that sense, the model is often the final layer rather than the foundation.
Nassau Street Partners deserves credit for placing those limitations at the centre of the conversation. A less serious event might have presented AI as an all-purpose answer. Instead, the firm used the summit to examine the boundary between automation and oversight. That approach is well suited to private markets, where confidence must be matched by discipline and where relationships still carry significant weight.
The broader message from London was that family offices are no longer peripheral sources of capital. They are increasingly active participants in the architecture of private finance. They can move across asset classes, take long-term positions and engage directly with companies and managers. As their role expands, the ability to combine technology with judgment will become a defining capability.
Their influence is also changing the way advisers and companies approach the market. A family office may bring more than funding. It may offer sector experience, a long holding period and access to a network built over generations. For businesses navigating uncertain exit markets, that form of aligned capital can be particularly attractive. The challenge for advisers is to understand the distinct objectives behind each pool of family wealth rather than treating it as a single category.
AI will not make every family office a better investor. It may even create new forms of false confidence. But used carefully, it can help small teams operate with institutional reach while preserving the independence that makes family capital distinctive. Nassau Street Partners’ summit captured that opportunity without ignoring the risks. In private markets, where the best decisions are rarely obvious and the best opportunities are rarely standardized, that balance may prove more valuable than any single model.