Behind private wealth sit confidential records, important relationships, critical systems and decisions that must remain sound over time. Exposure can arise not only from financial choices, but from how information is accessed, shared, retained and recovered.
I provide independent, executive-level judgement on technology, information security, data and AI for trustees, family offices and private clients. The aim is to make consequential decisions clearer and to strengthen control and continuity, working alongside established professional advisers.
Agron Fazliu · Independent technology adviser
In a family office, trust structure or private advisory business, even well-run arrangements depend on access rights, reliable records, clear decisions and providers who understand their responsibilities. The questions worth asking are often practical: Who can see sensitive information? What is the authoritative record? Can a critical process continue if a person or system becomes unavailable? What authority should an AI tool ever receive?
Entity records, mandates, documents and reporting may sit with different people or advisers. Reconciliation by spreadsheet feels manageable until a board, trustee or principal needs one answer that can be relied on.
Access is not confined to employees. Trustees, accountants, lawyers, custodians, outsourced administrators and temporary specialists may each hold a different piece of the same information. Permission needs to reflect the obligation and the purpose, not just convenience.
A small team can be highly capable and still depend on knowledge that has never been documented: where records live, how approvals work, which integration fails quietly and how to recover it.
Systems that work well can still create excessive dependence. A renewed contract, failed provider, changed data policy or difficult exit can expose how few choices remained.
Drafting, searching and summarising are useful. But an answer that reveals the wrong record, crosses an information boundary or cannot be checked is not an acceptable shortcut for fiduciary work.
Recovery should work when staff are absent, a provider is unavailable or a critical account is compromised. A documented plan is only a starting point; evidence comes from testing.
How these decisions are examined
An investment in technology may affect confidentiality, regulatory obligations, operational continuity, cost and the experience of the people who depend on it. Lawyers, fiduciaries, accountants, banks, internal teams and specialist providers each bring knowledge that matters. The challenge is to understand the combined implications before a decision becomes difficult to reverse.
My role is to make those implications visible. It does not replace the expertise of established advisers. It helps a principal, trustee, board or executive team ask sharper questions, test assumptions and decide what to approve, defer or change.
The questions can begin with something apparently narrow: a new provider, cloud system, AI assistant, access change or reporting problem. The useful assessment connects that immediate choice with the information it touches, the people who rely on it and the consequences if it fails.
Sometimes the answer is a different provider. Sometimes it is a smaller scope, stronger control, a better operating process, or no new system at all.
When client records, trust and entity information, approvals and outside service providers must remain consistent, controlled and retrievable. The objective is not to give technical advice on the trust itself; it is to make the systems and operating arrangements supporting fiduciary work more dependable.
When investment information, legal entities, administrators, banks, advisers and family stakeholders create a more complex information and supplier landscape than a small permanent team can comfortably oversee. The objective is clear reporting, resilient operations and decisions the office can defend later.
When personal devices, identities, communications, travel access, home technology and external advisers intersect with sensitive family and business information. The scope is assessment, architecture, supplier oversight and coordination with appropriate specialists, not personal protection or around-the-clock monitoring.
I also work with the lawyers, trustees and other trusted advisers who are often best placed to recognise when an independent technology view would help their client.
These are examples of the questions and reasoning I use. They are not client stories or claims about work carried out for a particular family, trust or institution.
Before buying another reporting platform, establish which figures are authoritative, who reconciles them, which exceptions matter and who owns corrections. A dashboard cannot fix disagreement about the underlying record.
Before discussing models, identify which people may see which documents, whether a result can be traced to its source, what the provider retains and which actions require approval. The design begins with the obligation, not the demo.
Ask what information and operating knowledge the provider alone holds, which alternative can replace it, how long a change would take and whether the current arrangement has been tested under disruption.
Who is accountable, what must be protected and what evidence may be needed later?
A faster system is not better if it is unclear which information is correct.
A useful process still needs the right boundaries between people, entities, advisers and automated tools.
Supplier exit, recoverability and transferable knowledge matter before they become urgent.
Define what better control, lower cost, usable reporting or reliable recovery would actually look like.
AI is often presented as a question of model capability. In a confidential organisation, it is first a question of authority: which records a system may read, what it may repeat, what it may draft, and which actions it may take without human approval. Those are different decisions with different controls.
The six levels of AI authorityI have held responsibility across technology leadership, information security and risk, data, systems and the executive decisions that connect them. I have worked on the operating questions that arise when confidential information and organisational obligations constrain a technology choice. The usefulness of that breadth is practical: I can question a proposal at both the management and technical level, without assuming the problem belongs to one discipline.
I do not publish employers, clients or identifiable assignments here. In confidential settings, a public list is often a poor way to demonstrate discretion. I prefer to show the questions I ask, the evidence I look for and the limits of the advice I give.
More about my approach
An initial conversation does not need confidential documents.
If a decision is approaching, a short outline is enough to see whether an independent view would be useful: the kind of organisation, the decision and the timing. Please do not include client identities, account details, private documents or other sensitive information in an initial LinkedIn message. If I am not the right person, I will say so.
