Every brokerage in New Zealand is being pitched artificial intelligence (AI) in some form. Few of the pitches mention the unglamorous prerequisite: none of it works without clean, consolidated client data. That is the lesson Nic Tyson (pictured), director of Taranaki-based broking firm The Advisers, took from four years of modernising a business he bought as a paper-based operation.
"A lot of what the AI journey requires is a true source of data - something our disjointed system didn't have," Tyson said.
When Tyson acquired the firm, its client knowledge was scattered across filing cabinets and a collection of software tools that each held a fragment of the picture but never spoke to one another. Reconstructing a client relationship meant manual detective work. Before any AI tool could be useful, that fragmentation had to go – so the firm consolidated its records and documents into a single cloud-based system, JAVLN's Officetech platform.
Tyson's sequencing - consolidate first, automate second - matters because AI tools inherit the quality of whatever data sits beneath them. A summary generated from incomplete or contradictory records is worse than no summary at all. Getting the records right is not only an AI enabler; it is a licensing obligation. Standard Condition 1 of every financial advice provider (FAP) licence requires firms to create records in a timely manner and maintain adequate records of their advice service, and the Financial Markets Authority (FMA) has said its monitoring of the previous advice regime consistently identified poor record keeping as an area of concern. A true source of data, in other words, is where the commercial and regulatory arguments converge.
The direction of travel is more scrutiny, not less. The Conduct of Financial Institutions (CoFI) regime has been in full effect since 31 March 2025, with the FMA taking an outcomes-focused approach to how consumers are treated, and the Contracts of Insurance Act 2024 - described by the regulator as a significant reform of New Zealand's insurance contract law - comes into effect on November 15 2027. Each layer of that framework ultimately asks the same question of a brokerage: can you show, quickly and completely, what happened with this client?
Tyson said the consolidation decision came out of due diligence on where the technology was heading, rather than any single feature - the firm could see the platform's roadmap and wanted the foundation in place before the AI capabilities arrived. Being early, he said, has added considerable value.
The payoff arrives as narrative. Ask the system about a client and it assembles the file notes, correspondence, documents and tasks into a coherent account of the relationship.
"The tool gives you the story of a client pretty quickly - you run the report and it comes back to you fast," Tyson said.
That capability produced benefits Tyson admits he never anticipated when he signed up - oversight of what is happening across his advising force, not just within individual client files. He believes the value scales with size: the bigger the firm, the harder it becomes to maintain a view of everything, which is where a consolidated record earns its keep. Notably, the brokerage does not use the technology for underwriting-style decisions about which clients to take on; for The Advisers, it is squarely about understanding what has transpired with a client.
None of this settles the wider questions about AI in financial services - the FMA's Financial Conduct Report 2026 lists financial advice among its regulatory priorities and supervisory expectations will keep evolving. But Tyson's experience suggests the first move is the same regardless of which AI tools a brokerage eventually chooses: build the single, reliable record first. The intelligence is only ever as good as the data it reads.
