At the Hubbis Wealth Planning and Structuring Forum Hong Kong 2025, Ada Dong, Vice President of Sales at Eton Solutions, set out a pragmatic blueprint for bringing artificial intelligence into family office operations. Her case was grounded in enterprise-grade security, disciplined data architecture and practical workflows that reduce reconciliation, document handling and reporting effort. Dong also highlighted Eton’s ISO 42001 certification for responsible AI and a single-tenant cloud approach designed for regulated environments.
Key Takeaways
- Infrastructure Before Intelligence: Many Asian family offices begin with elite advisers and sophisticated trust planning, yet operate day-to-day on spreadsheets and fragmented systems. Dong emphasised that AI can only add value once core data architecture, accounting systems and reporting frameworks are consolidated and automated.
- Integrated Platforms Enable AI Impact: Eton positions full-stack infrastructure as the prerequisite for automation. By unifying performance reporting, general-ledger accounting and tax ledgers, families can achieve a single source of truth that supports multi-entity, multi-generation visibility and eliminates reconciliation bottlenecks.
- AI Success Depends on Data Discipline: Referencing recent research, Dong noted that according to an MIT study, ninety-five per cent of enterprise AI initiatives fail due to various reasons, especially due to unstructured, inaccessible or siloed internal data, where AI agents cannot extract value from PDFs, emails and scattered files. Families must convert documents into structured and vector databases to unlock meaningful automation.
- Secure, Responsible and Private AI Deployment: Dong highlighted that bringing AI to the client’s private cloud, rather than sending data externally, is essential for regulated wealth environments. Eton’s ISO 42001 responsible-AI certification, alongside SOC and ISO security standards, underpins auditability, privacy and governance.
- Augmentation, Not Replacement: AI is designed to remove repetitive administrative tasks, not replace talent. Dong stressed that future-ready family offices will be those where analysts and advisers use AI as a productivity partner, enabling deeper focus on strategic decisions, family governance and investment oversight.
Dong began with a challenge that many Asian families face: they set up family offices with the best advisers in trust structuring, legacy planning and residency strategy, only to discover that day-to-day operations rely on spreadsheets and manual information gathering. This “advice first, infrastructure later” reality means principals struggle to see a consolidated picture of holdings, cash positions and private market exposure across banks, funds and direct deals. Juniors spend most of their time collecting statements, logging into portals and updating Excel files, creating talent drain and operational fragility.
Family offices typically evolve through distinct stages: ad-hoc spreadsheets, isolated performance or accounting tools, and eventually integrated platforms. Eton positions itself at the fully-integrated end of this spectrum, combining performance reporting, general ledger accounting and tax ledgers in a single system. Dong described this as the foundation required before AI can deliver meaningful impact, since reconciliation must be automated and data must be accurate, structured and controlled.
Configuring for Multi-Generation, Multi-Entity Wealth
Dong emphasised the importance of a configurable architecture for modern wealth structures. Asian families increasingly operate through multiple trusts, family-office vehicles, venture-capital entities and co-investment arrangements. Many have shifted from being limited partners to acting as general partners in their own funds or direct-deal structures.
Eton’s platform creates a look-through view down to each sub-fund, property and marketable security, allocating proportional ownership across family members and external co-investors. Permissions align to governance rights, enabling principals, heirs and investment partners to see only what they own. A client portal supports transparency for friends-and-family co-investors, without compromising data secrecy.
Dong noted that this is becoming essential as wealth becomes institutionalised within families. “It takes a village” applies not only to trust administration but also to modern private-wealth operations, where accounting, investment, tax, legal and administrative functions intersect.
Why Many AI Initiatives Fail
Dong referenced a recent Massachusetts Institute of Technology study showing that ninety-five per cent of organisations investing in AI see no productivity gains. The core reason is not AI model’s capability but enterprises’ data readiness. Large models can complete tasks, but they cannot interact with unstructured, inaccessible or siloed data. Folder-based storage, scattered PDFs and handwritten materials render AI ineffective.
Department-level tools, such as standalone copilots, create inconsistent answers and workflow clashes between teams. The result is duplicated work, mistrust in outputs and added friction rather than efficiency. For family offices handling sensitive financial data, hosted single-point AI tools also raise material confidentiality and sovereignty concerns.
Bringing AI to the Data, Not Data to the AI
Dong outlined an alternative approach. Each family client receives its own isolated cloud instance, and the AI module is deployed to that environment rather than sending data to external providers. This preserves confidentiality and control while allowing models to interact directly with the source data.
Eton helps clients convert unstructured materials into structured databases, then into vector databases, enabling search, retrieval and automated processing. With this foundation, AI agents can read, classify and extract information, generate entries and route documents into the correct workflows with audit trails.
Dong stressed that this architecture is designed to align with regulated-industry expectations: secure, auditable, explainable and tailored to control hierarchies. In her framing, AI does not replace roles; it elevates them by removing repetitive work and retaining institutional knowledge.
From Technology to Tangible Outcomes
Dong transitioned from architecture to outcomes. While AI is often discussed in abstract terms, she focused on real operational gains inside family offices. Document processing remains one of the most resource-intensive tasks, particularly in Asia where physical statements, contracts and handwritten notes are still common. AI agents can ingest PDFs, scanned documents and emails, extract relevant fields and populate systems or CSV files without manual input.
Monthly reconciliations, capital-call notices, private-equity statements and property-management records often consume the bulk of junior analysts’ time. Dong explained that automated extraction and categorisation allow teams to shift attention from checking data to making decisions based on it. She observed that “the low-hanging fruit is not in sales and investment, but in operations”, where structured automation drives immediate productivity improvements and supports retention by reducing monotonous work.
An AI-Augmented Office, Not an AI-Replaced One
Dong reiterated a reassuring and practical message: AI is not replacing family-office talent, it is replacing repetitive tasks that drain time and motivation. She noted that labour markets are tightening and competition for skilled hires is intensifying. In her words, “AI will not replace workers, but workers who do not know how to use AI will be replaced.” The implication is clear. AI fluency will become an expected competency in modern wealth-office environments, particularly as multi-asset portfolios and cross-border structures grow in complexity.
The goal, she explained, is to create an AI-augmented team in which analysts, accountants and investment staff operate with real-time context, workflow support and intelligent automation embedded into everyday tools, from portfolio dashboards and reporting systems to communication platforms like Outlook and WhatsApp.
Responsible AI and Enterprise-Grade Standards
Eton Solutions has positioned responsible AI at the core of its value proposition. Dong highlighted the firm’s ISO 42001 certification for responsible AI, alongside ISO 27001 for security, ISO 27701 for privacy and SOC 1 and SOC 2 attestations. This governance layer differentiates enterprise-grade AI deployment from consumer models.
This discipline is critical for wealth environments, where confidentiality, auditability and model explainability cannot be compromised. Dong pointed out that audited decision trails ensure traceability, and isolated client cloud environments eliminate data co-mingling risk. Standards, she argued, matter as much as technology itself in a regulated industry.
The Path Forward
Dong closed by reinforcing the strategic imperative for family offices to digitise intelligently, not reactively. AI adoption is not a single project but a phased journey grounded in data readiness, operational clarity and governance discipline. The opportunity is not limited to efficiency. As families evolve into institutional wealth managers and co-investors, precision, transparency and scalability become strategic assets.
Her final message to the audience was pragmatic and optimistic: family offices that adopt structured, secure and intentional AI frameworks will build resilient infrastructure capable of supporting multi-generational wealth. Those who overlook the operational layer, she noted, risk bottlenecks that undermine even the most sophisticated legacy structures.
Source: https://www.hubbis.com/article/adapting-to-tomorrow-how-ai-can-future-proof-the-family-office