The reaction to Anthropic's launch of Claude for Financial Advisors has been fun to watch.
Some incumbents immediately explained why Claude won't replace their platform. New AI firms rushed to explain why their approach is superior. A debate quickly started around who has the best AI for financial services.
But I think that what Anthropic and its ecosystem partners have done is give us a glimpse of what the future wealth management operating model is likely to look like, in part.
At first glance, the announcement looks like an AI story. Claude can connect to custodians, portfolio systems, CRMs and planning tools, helping advisors prepare for meetings, analyse portfolios and automate administrative work.
But look more closely and it becomes something else - an orchestration story.
Anthropic isn't threatening to replace the systems wealth managers are using today, but instead it’ll sit above them, coordinating them, pulling information from multiple sources and creating a layer of intelligence that can operate across the entire estate.
That's important because, for decades we have built wealth management technology vertically. Every new problem produced a new platform. Every new platform created another data source. Every integration promised a unified picture but usually delivered a partial one.
AI changes the economics of that fragmentation. When a human advisor encounters fragmented data, it's inefficient and unscalable, but when an AI encounters fragmented data, it's strategically limiting.
The quality of the AI output becomes directly linked to the quality and completeness of the underlying data. Context is key.
In that sense, the announcement is really about something much bigger than Claude: the emergence of data orchestration as a competitive capability.
The firms that can bring data together across custodians, portfolio systems, planning tools, CRMs and proprietary sources will have a fundamentally different AI experience from those that can’t. They won't simply have better answers, they'll have richer context, greater consistency and an increasingly proprietary understanding of their clients and their business.
That data layer becomes strategic because it compounds over time. Every system connected to it strengthens it. Every workflow built upon it increases its value. Every interaction enriches it further.
In a world where AI capabilities are becoming increasingly accessible (and commoditised?), the data foundation becomes the moat for the manager.
But this data orchestration challenge is only one half of what's emerging through the Anthropic announcement. The second is one the industry is only beginning to think about: AI orchestration.
Much of today's discussion assumes firms need to back a winner. Claude, OpenAI, Gemini, Microsoft or whichever model currently occupies the top of the leaderboard.
But the pace of development is simply too fast. Models improve. Relative strengths shift. New entrants appear. Regulatory requirements evolve. Costs change. Decisions made in an unpredictable global climate, influence technology supply chains.
The strategic objective needs to shift focus to retaining the freedom to use whichever model is right for a given task.
The most sophisticated wealth managers will almost certainly end up operating multiple forms of AI simultaneously. Different models, agents and specialised services will perform different functions across research, operations, reporting, client servicing and investment decision support.
The challenge therefore isn't selecting AI. And that's where I think Anthropic's announcement is both important and incomplete.
The connectors solve part of the problem. They connect intelligence to data. What they don't yet solve is how wealth managers retain ownership of that relationship.
How do firms maintain a complete view of their data independent of any custodian? How do they switch AI providers without rebuilding workflows? How do they allow multiple agents to collaborate against the same business context? How do they avoid replacing platform lock-in with AI lock-in?
Those questions will matter far more over the next decade than whether Claude is marginally better than its competitors this year.
Because ultimately, data orchestration and AI orchestration are solving the same problem.
Control.
Control of data.
Control of intelligence.
Control of strategic direction.
The wealth management industry has spent years wrestling with dependency on platforms, vendors and ecosystems. The opportunity now is to build an architecture that allows firms to benefit from all of them without becoming captive to any of them.