AI has quickly become the dominant theme in wealth management technology discussions. But when wealth managers, advisers, platform providers and technology leaders sit down to talk honestly about it, the conversation tends to drift somewhere less glamorous: legacy systems, fragmented data and the technical debt that has built up over years of growth.
That is exactly what happened at Owen James’ WealthTech Matters event in London on 8 October 2026, which brought together COOs, CTOs and CIOs from wealth managers, private banks, DFMs, advisers and multi-family offices. FA Solutions was a headline sponsor, and I led our roundtable, Connected by Design: Building a Technology Ecosystem That Evolves With Your Business, facilitated by Sharmil Patwa, founder of Opus Una. The session was held under Chatham House rules, so the views shared here are not attributed to individuals or firms.
Are we building on a broken table?
The question that framed the roundtable discussion
This article draws on that discussion, along with my wider conversations across the industry. My conclusion is that many firms are approaching AI from the wrong direction. The firms most likely to benefit over the long term are not necessarily those making the biggest AI investments today. More often, they are the ones quietly modernising their operating models, improving data quality and building architectures flexible enough to change as the market does.
In other words, the real differentiator may not be AI itself. It may be the quality of the foundations underneath it.
AI Is Being Treated as the Solution Before the Problem Is Fixed
It’s easy to see why AI has captured so much attention. Boards are asking for AI strategies, vendors are building AI into their product roadmaps, and advisers are exploring how automation can improve productivity and client servicing. Everywhere you look, the conversation is about what AI might enable.
The pressure behind that conversation is real. Wealth management is entering a period of significant change:
- Operating models are expected to shift. Research by ThoughtLab, Deloitte and FNZ found that 69% of wealth management executives believe AI will significantly change the way their firms work.
- Client expectations are rising. The same research found that 68% of investors want digital experiences that match those of leading born-digital companies.
- Cost and regulation aren’t going away. Firms still need to improve efficiency, reduce operating costs and respond to an evolving regulatory environment.
- Sources: Deloitte – The evolution of wealth management through AI, Businesswire)
The challenge is that AI is often being positioned as the answer before firms have fully addressed the problems underneath it. And as our roundtable made clear, those problems are usually structural rather than technological.
AI Doesn’t Fix Fragmented Data. It Amplifies It
Most wealth firms have built up their technology over many years. New systems were added to meet changing client needs, regulatory obligations, business growth and acquisitions, and each investment made sense at the time. The result is that many firms now run complex estates where:
- client data is spread across multiple platforms,
- business processes cut across numerous systems,
- manual intervention is still an essential part of day-to-day operations.
Firm consolidation can magnify this. Participants in our roundtable described how, after an acquisition, a firm can end up with similar capabilities delivered through entirely different tools, which makes a full technology review unavoidable. The group named data fragmentation, duplicate data and inconsistent conventions as some of the biggest barriers to both integration and AI.
Historically, firms could live with this. People developed workarounds, teams learned where information sat, and operational expertise bridged the gaps between systems.
AI changes that. It depends entirely on the quality, consistency and accessibility of the data it can reach. If data is fragmented, duplicated or hard to retrieve, AI doesn’t remove those issues. It magnifies them.
More than 90% of data users in banking report difficulty accessing the information they need, and 81% identify data quality as a major challenge.
Source: Deloitte
These figures come from banking, but the issue extends well into wealth management. The roundtable’s consensus was clear: treat your core systems as systems of record first. Establish what data you need, how it is used and whether it can be extracted, before building any workflow or AI agent layer on top.
Four Questions to Answer Before Scaling AI
Before firms can implement AI at scale, they need answers to some more fundamental questions:
Is your data ready for AI?
- Where does critical client data reside?
- Which systems act as the source of truth?
- How easily can information move across the organisation?
- Who owns the data, and how is it governed?
These questions are less exciting than generative AI or autonomous agents. But they are the ones that decide whether AI initiatives deliver meaningful outcomes or stay trapped in pilot programmes.
Beyond Platform vs Best-of-Breed: Connected by Design
These same considerations are reshaping one of the industry’s longest-running debates. Firms have traditionally had to choose between:
- an integrated platform that covers most required functions, but may be less sophisticated in some areas, or
- a collection of best-of-breed tools that go deeper in individual areas but are harder to join up.
Increasingly, neither option fully reflects how modern firms want to operate. Technology is moving too quickly to assume that today’s preferred solution will still be the right answer in five years’ time.
What participants want from their ecosystem
The roundtable gave a clear picture of what this looks like in practice:
- An orchestration layer that holds it all together. Participants saw the process and orchestration layer as the most important part of the estate, whether it is managed manually or through technology. One firm had tackled large legacy systems by building a process-management layer above them rather than replacing them.
- A single adviser-facing application. The ideal future state was one application that aggregates data and orchestrates the full adviser journey.
- Components that can be swapped without disruption. A stable orchestration layer would let firms replace individual tools without breaking the wider process.
- Buy before build, where possible. One firm stated a clear preference to buy unless a capability simply can’t be purchased, recognising that proprietary development creates ongoing maintenance and technology debt. Others stressed that both approaches remain viable, depending on a firm’s capability and appetite for ownership.
The right architecture also depends on the business model. A front-to-back wealth manager may need integration with custodians, market-data providers, CRM, accounting and messaging standards. An adviser business may prioritise KYC, suitability and prospecting tools instead.
Having an API Is No Longer Enough
Adaptability matters because the pace of change goes well beyond AI. Client expectations, regulation and technology will all keep moving, and firms need estates that can respond without repeated large-scale transformation programmes. That’s why APIs and interoperability have moved out of the IT department and into the boardroom.
The existence of an API is no longer enough. Leaders want to know how open their ecosystem really is. My view is that API-first design has to be built into a product from the start, not patched on later. Participants agreed that API claims should be tested for meaningful coverage: an interface that exposes only a small share of data or functionality can’t support a genuine ecosystem. Integration platforms can help fill gaps, but for strategic platforms, complete native access matters more.
Contractual commitments are also shifting, driven by the pace of change. Participants distinguished between strategic backbone systems, where longer commitments can make sense, and peripheral plug-and-play services that need more flexible terms. But flexibility has a limit: advisers can resist repeated changes to the tools they use every day, so user adoption needs to be part of any replacement decision.
Questions to ask about any platform in your ecosystem
- Can new capabilities be integrated quickly?
- Can data move seamlessly between platforms?
- How much of the platform’s data and functionality does the API actually expose?
- Can future AI services access the information they need without extensive redevelopment?
- Do you own your data, and do you hold a copy of it?
- What do your contracts say about exit, data extraction and portability?
- How much influence do you have over the supplier’s product roadmap?
Governance Is What Allows AI to Scale
There’s often an assumption that governance slows innovation, particularly where AI is involved. In reality, strong governance is what allows innovation to scale safely and sustainably.
The FCA has said it does not plan to introduce AI-specific rules. Instead, it will rely on existing frameworks such as the Consumer Duty and the Senior Managers and Certification Regime, noting that its rules on accountability for senior managers are relevant to the safe use of AI. Existing governance frameworks remain highly relevant, even as the underlying technology changes. (Sources: AI and the FCA, AI in financial services)
Roundtable participants broadly agreed that governance is necessary. They also warned that traditional, risk-averse approaches can hold innovation back when compliance and risk teams lack AI expertise. To close that gap, they pointed to:
- Clear accountability for AI-assisted outcomes.
- Detailed evidence, including audit logs, records of AI activity, human sign-off and decision trails, so firms can demonstrate control if the regulator asks.
- Visibility and control over what AI can access, where data resides, whether it is segregated, and what monitoring, security and audit controls apply.
- Dedicated capability, such as technology-risk or AI-focused specialists.
What the Right Foundations Make Possible
Clean data and connected systems are not the end goal. They are what allow firms to change the advice proposition itself. The roundtable discussed a shift from periodic, reactive servicing to proactive engagement, where firms identify client needs and step in throughout the relationship.
Participants also explored automated compliance checks on every client file rather than a random sample, and the possibility that lower-value advice segments could increasingly be served through targeted support, recommendations and digital nudges rather than a traditional annual review. These were raised as possibilities rather than conclusions, which is exactly why architecture needs to absorb them without another wholesale replacement.
The Bottom Line: Foundations First
The firms likely to gain the most value from AI will not be those with the most ambitious AI vision. They will be the firms that have invested in clean data, connected systems, sound governance and adaptable architectures. While much of the industry looks for the next breakthrough technology, these firms are focused on something less glamorous but arguably more important: creating the conditions that allow future innovation to succeed.
The future of wealth management will undoubtedly involve AI. But before firms ask what AI can do for them, they should ask a simpler question:
Is your technology foundation ready for what’s coming next?
The future of wealth management won’t be defined by who adopts AI first. It will be defined by who builds an organisation capable of adapting fastest. That journey starts long before AI. It starts with the foundations.