What Is AI Paraplanning? An Australian Adviser's Guide

AI paraplanning is the most misused phrase in Australian advice tech right now. Vendors apply it to anything with a language model attached, which means an adviser comparing two tools is often comparing a meeting transcriber to a strategy engine and being told they do the same job. They do not.

So here is the clean definition, the one worth keeping.

AI paraplanning is software that automates the core tasks of a paraplanner: research and data collection, strategy modelling and comparison, and the production of draft advice documents, while the licensed adviser remains responsible for the advice itself.

That sentence does the work. The rest of this piece defends each part of it, because the boundaries are where firms waste money.

AI paraplanning vs AI file-noting vs AI advice generation

These are three different layers. Most tools sold as "AI for advice" sit in exactly one of them, and the category they occupy determines what you should pay and what you can expect.

File-noting is the narrow case. It captures what was said and shapes it into a record. Useful, and overlapping with generic transcription tools, but it stops before strategy. We cover that boundary in detail in generic AI note-takers vs adviser-specific AI and in ASIC-compliant file notes.

AI advice generation is the broad case. It runs the whole production line from fact-find to finished document inside one governed environment.

AI paraplanning is the layer in the middle, and it is the layer where most of the time savings in the category actually come from, because paraplanning is the bottleneck in most firms. Mix the three up and you will buy a transcriber when you needed a strategy engine, or pay for an advice OS when a file-note tool would have done.

What AI paraplanning actually does

Three tasks. Break them out, because a tool that does one well and the others poorly is common.

Research and data collection

The AI pulls together the inputs a paraplanner would otherwise assemble by hand: current super and investment positions, insurance held, fees, product features, contribution caps, and the relevant thresholds for the strategy under consideration. It reads the file note and the fact-find and assembles the picture. The adviser checks it. This is where the first hour per file usually goes, and it is the most mechanical hour.

Strategy modelling and comparison

The AI runs the numbers on the options: contribution strategies, transition-to-retirement structures, debt recycling, insurance restructures, and projects the outcomes side by side. It does not decide which strategy is right. It does the arithmetic and surfaces the comparison so the adviser can exercise judgement faster.

Draft document production

The AI writes the first draft of the advice document, the SOA or the record that replaces it, populated with the client's data, the modelled strategy, and the reasoning. The adviser edits, challenges, and signs. A first draft is not a final document, and treating it as one is the single most common way firms get AI paraplanning wrong. For how this plays out specifically in SOA production, see SOA automation and AI statement-of-advice software.

A worked example: transition to retirement

Take a 61-year-old client wanting to reduce hours and start drawing down super. Here is who does what.

1. The adviser runs the client meeting and confirms the goal, risk profile, and constraints. The file note captures it.
2. The AI reads the file note and fact-find, pulls the client's super balance, current contributions, and the relevant preservation and tax settings, and assembles the data pack.
3. The AI models the transition-to-retirement pension alongside a salary-sacrifice arrangement, projects the after-tax income and the super balance at 65 under each, and lays the two options side by side.
4. The adviser reviews the modelling, applies judgement about the client's circumstances the model cannot see, and selects the strategy.
5. The AI drafts the advice document with the chosen strategy, the projections, the product comparison, and the reasoning.
6. The adviser edits and signs. The advice is theirs.
7. The licensee sees the full chain: the inputs that went in, the modelling that ran, the draft the AI produced, and the adviser's changes before sign-off.

The work that compressed is steps 2, 3, and 5. The judgement, steps 1, 4, and 6, did not move. That is the whole point of the layer.

The responsibility line: who is accountable for what

This is the part vendors skip and the part ASIC cares about most.

The AI is not regulated. The advice is. ASIC's Regulatory Guide 175 places the obligation to provide appropriate advice on the licensed adviser, regardless of which tools produced the draft. No model output carries that responsibility. The adviser does.

That has two practical consequences for any tool you evaluate.

First, the adviser has to remain the decision-maker in the workflow, not a rubber stamp at the end of an automated pipeline. A tool that makes it easy to ship an unread draft is a governance problem wearing a productivity badge.

Second, the audit trail has to be complete. Under ASIC's record-keeping expectations in RG 36, the licensee should be able to reconstruct, for any piece of advice, what inputs went in and what the AI produced before the adviser touched it. One chain, not three vendor exports stitched together after the fact.

If you want a thirty-minute exercise: take your last five completed SOAs and ask whether you could show a reviewer the exact inputs and the AI's first draft for each. If you cannot, that gap is your real AI paraplanning requirement, ahead of any feature list.

Why the Australian context matters

A US or UK paraplanning tool can be excellent and still be wrong for an Australian licensee, for two reasons.

Data residency is the first. Client PII for advice is sensitive, and a growing number of licensees now require written confirmation of where the data sits and where the model inference runs. Onshore is increasingly a procurement filter, not a preference. Ask the question in writing.

The regulatory direction is the second. The Delivering Better Financial Outcomes reforms propose replacing the SOA with a shorter Client Advice Record that can be a video or interactive format rather than a written document. Tranche 2 stalled through 2026 in the aftermath of the Shield and First Guardian collapses, and the minister has set no date. The direction is clear regardless: the document gets shorter and more flexible. A paraplanning tool built around the old fixed SOA template will be modelling the wrong artefact when that lands. Build around the strategy and the record, not the format.

This shift is happening against a profession that keeps shrinking. Adviser numbers sat at roughly 15,142 in March 2026 on Padua Wealth Data's analysis of the Financial Adviser Register, down about 429 over the prior twelve months. Fewer advisers, the same demand, and the same compliance load. AI paraplanning is the lever most firms reach for first, which is exactly why getting the definition and the boundaries right matters before you spend.

Frequently asked questions

Is AI paraplanning ASIC-compliant?

The tool is neither compliant nor non-compliant on its own. The advice is what gets regulated, and ASIC RG 175 keeps the responsibility with the licensed adviser. A compliant workflow keeps the adviser as the decision-maker and maintains a complete audit trail. The tool supports that. It does not replace it.

Can AI paraplanning replace a human paraplanner?

It replaces the most mechanical parts of the role: data assembly, first-pass modelling, and draft production. It does not replace the judgement a good paraplanner applies, and it does not carry the adviser's accountability. Most firms redeploy paraplanner time toward checking and higher-value work rather than cutting the role.

Is AI paraplanning the same as AI file-noting?

No. File-noting turns a meeting into a structured note and stops there. Paraplanning takes the note and fact-find and does research, modelling, and draft-document production. A file-note tool will not produce your SOA.

Who owns the output the AI produces?

Commercially, check the vendor's terms, because they vary. Professionally, the adviser owns the advice the moment they sign it, regardless of what drafted the document. That is the line that matters for accountability.

Does it work with XPlan and Iress?

It depends on the tool. Some AI paraplanning tools integrate with XPlan and other practice software; others expect to run alongside or replace parts of it. Ask specifically how data moves in and out before you assume integration.

What happens when the AI gets something wrong?

The adviser catches it at review, which is why the review step is not optional. The risk is not that the AI errs. It is that a firm builds a workflow that ships drafts without a real check. Design the check in.

What does AI paraplanning cost?

Pricing transparency is uneven across the Australian category. Some tools publish pricing, most require a sales conversation. If transparent pricing matters to your procurement, factor the opacity itself into the comparison.

Where is my client data hosted?

For some tools, onshore in Australia. For many, US or global cloud regions by default. For both PII storage and the model-inference layer, ask for the data region in writing. For an Australian licensee, this is a procurement question, not a technical footnote.

This article is general information about software for the Australian advice profession. It is not personal advice and does not consider any firm's specific circumstances. Conduct your own due diligence on each vendor's current capabilities, data residency, and governance posture.