|
Matthias Knab, Opalesque for New Managers: I spent two days recently at a closed-door gathering of family offices - single and multi-family, first generation to fourth, from Southeast Asia to the Nordics to the Americas. As always, the real value wasn't in any single presentation but in the pattern that emerged when you listened across all of them. Four themes kept resurfacing, and together they sketch a picture of where sophisticated private capital is heading - and where it is quietly worried.
1. The pendulum on direct investing has stopped swinging - and settled somewhere honest
For a decade, "going direct" was the badge of the ambitious family office. What I heard now was something more mature: a candid accounting of what direct investing actually costs, and who should be doing it.
The families doing it well share one trait: they are operators first, allocators second. One principal described himself as "an operator masquerading as a financial person" - for him, the diligence and the wire transfer are merely the beginning; the real work, and the real edge, starts when you roll up your sleeves inside the company. His family stayed within a tight thesis for six years and only one home market before expanding abroad, and even then only through local partners. The few times they were tempted outside their circle of competence, he admitted, "it never went well."
Equally striking was the honesty about measurement. Several speakers openly questioned IRR as the yardstick - too easy to engineer, too dependent on the flavor of the month. One family evaluates its direct book purely on a money-multiple basis over five, eight, ten years, accepting that in a large enough portfolio a few failures are the price of the outliers. Another participant put it plainly: for illiquid family holdings, "valuation is an art, not a science" - so pretending to precision is self-deception.
And private equity? Far from being dismissed, it was repositioned. The direct investors in the room see PE funds as their natural downstream acquirers - taking on much safer, later-stage risk - and as a vital release valve in what one speaker memorably called the industry's "exit desert." The relationship is symbiotic, not competitive. The consensus rule of thumb: if you go direct, you must expect to beat what PE would have delivered, because you are giving up diversification to get there. If you can't credibly clear that bar, be an LP.
2. Risk management is becoming a family discipline, not just a portfolio discipline
A panel nominally about black swans turned into something more interesting: a conversation about which risks families actually control.
One next-generation panelist had started his career at a storied institution that collapsed weeks after he joined. His dry observation - when there are TV cameras outside your office, the warning signs are probably there - drew laughter, but his point was serious: employees and insiders were told until the last minute that rescue was coming. The lesson for families: institutional reassurance is maybe not what it seems.
From the shipping side of the room came the most quotable risk philosophy of the event. In a deeply cyclical, derived-demand industry, the discipline is to position deliberately rather than place bets, never follow the crowd, and keep patient capital ready to be the buyer on the dip - not when everyone else is piling in and rates are sky-high. That is portable wisdom for any asset class in 2026.
But the sharpest insight concerned succession. One panelist recalled a line she had heard elsewhere: the problem with family office transitions is that everyone focuses on the office part and not the family part. Structures, holdcos and investment committees get endless attention; trust across generations gets almost none. Her framework for the next gen: know when to honor what was built, and when to refine direction - because the heir who changes nothing and the heir who changes everything out of ego are equally likely to destroy the enterprise.
And on reputational risk, a contrarian note I found genuinely useful: the greater danger for most families is overreacting, not underreacting. Much of what looks like a reputational crisis is noise with no material impact on stakeholder trust. Map your vulnerabilities in advance, understand which ones actually matter, and respond proportionately. Corollary: the less information about the family that sits online, the less raw material exists to be misconstrued later.
3. AI has moved from panel-filler to working infrastructure - with a surprisingly sober playbook
Two years ago, AI sessions at family office events were vision statements. This one was a tooling discussion. Families on stage described multi-agent systems already running: an in-house AI agent trained on public market data supporting the liquid-strategies team; agents drafting research with citations back to source documents so a human can verify before deciding; even a "digital twin" one principal is building to capture and transfer her own knowledge - contacts, patterns, forgotten details - across team members and generations.
Three principles ran through the discussion, and I'd endorse all of them:
First, governance before tools. Without a strategy and adoption framework, "the best tool will not find you any solutions." The tool landscape shifts weekly; what matters is knowing your own processes and use cases.
Second, know what the machine is actually doing. One panelist gave the cleanest explanation of hallucination I've heard at an investor event: the model isn't lying or being creative - it's computing probabilities, and where its data is thin, the highest-probability answer is still a low-probability answer that merely looks random to us. Hence the practical rule: AI is superb at pattern-recognition over material you give it - summarizing a data room, flagging red-flag clauses in legal documents - and still unsuitable for making the investment decision itself. Access to deals, several agreed, remains stubbornly human: it lives in networks, reputation and bargaining power, none of which an agent can yet manufacture.
Third, the "why bother" question got the best answer of the session. Family offices aren't competing for market share, so why adopt AI at all? The reply: if you still hold operating businesses, they will be disrupted whether you engage or not. And if you've sold and merely preserve wealth, taxes and inheritance mean that a family that doesn't grow is a family that shrinks - just slowly enough not to notice. Or, as the closing quip went: we all use phones and computers - these are tools. Why not this one?
4. The scarcest asset in the room was authenticity - and everyone knew it
The closing keynote, from a serial founder turned family principal, opened with an exercise: close your eyes and answer, what will your legacy be, and what matters most to you right now? Nearly every hand in the room, he predicted correctly, pointed to the same answer: family. His argument - delivered with the credibility of someone who had built and exited repeatedly - was that the 30-second LinkedIn pitch we all perform at these events is precisely what prevents the connections we came for, and that beyond a certain threshold, more money doesn't improve a life and frequently damages it.
But this wasn't a soft talk. His hard-edged thesis: the passive LP is an endangered species. His family will no longer fund a company unless several of its own network partners commit to becoming actual customers of it. He argued for pooling deal flow among trusted families rather than sourcing externally, for building intelligence networks that surface change before it happens, and for connecting private portfolio data across aligned investors while everyone else fights over public market signals. His summary line stuck with me: don't just build a fund - build a dynasty. Stop settling for double-digit-percentage thinking in a power-law world.
What ties it together
Across very different sessions, one thread: the professionalization of the personal. Direct investing works when the family's operational identity is real. Risk management works when the family - not just the office - is governed. AI works when humans stay in the loop and relationships stay human. And networks work when people stop performing and start connecting.
The families that will compound across the next generation are the ones treating trust - inside the family, across the portfolio, and among each other - as the asset class it actually is.
|
RSS
