Using AI Tools for Portfolio Analysis? Here's How to Disclose It

15 October 2026·5 min read

Quick answer: Yes, you can use AI tools for research summaries, scenario modelling and portfolio analysis — but if an AI-generated output genuinely shaped a client's outcome or recommendation, that use deserves a clear, plain-English disclosure a client would actually notice, not a buried clause in your FSG. AI tools are genuinely strong at summarising, comparing and drafting first passes; they're not a substitute for an adviser's judgement, and they can produce confident, polished, and occasionally wrong output that's easy to miss without a proper review step. 🚀

Here's the uncomfortable truth: most advice practices are already using some form of AI tool — for research summaries, meeting notes, first-draft commentary, even portfolio scenario testing — and most of them haven't said a word about it to clients. That's not necessarily dishonest, but it is a trust gap waiting to open, because clients are increasingly AI-literate themselves and will eventually ask. Being upfront about how and where AI fits into your process, before a client asks "did a robot write this," is a far better position to be in. 💖

What most advisers get wrong

There are two opposite failure modes here, and both are common. Some practices say nothing at all, assuming AI tools are just "research software" that doesn't need mentioning — fine until a client asks directly and the silence reads as evasive. Others swing the other way and bolt on dense, generic AI disclaimers that make clients picture their entire portfolio being run by an algorithm, which is rarely true and rarely helpful. The honest middle ground is specific: say what the tool actually does, and say clearly that a human adviser reviews and owns the outcome.

Please note: general information, not compliance or legal advice regarding your AFSL or licensee obligations — check current regulatory guidance (including from ASIC and your licensee) before relying on any disclosure wording.

The AI Disclosure Line-Up (copy, adapt, run past your licensee):

Website / "our process" page: "We use AI-assisted tools to help summarise research and model scenarios faster — every output is reviewed, checked and owned by your adviser before it ever reaches you."

SOA / client documentation footer: "Portions of the research and analysis supporting this advice were prepared with the assistance of AI-based tools, under the direct review and professional judgement of your adviser."

Verbal, in a client meeting: "We've started using an AI tool to help us crunch scenarios faster — it doesn't make any decisions, it just speeds up the number-crunching so we spend more time talking through what it actually means for you."

Internal rule of thumb: if an AI tool touched anything that shaped a recommendation, disclose it somewhere the client will see. If it only helped draft an internal meeting agenda, it probably doesn't need a client-facing mention.

Solo adviser practice: Uses an AI tool to summarise long fund PDS documents and research notes before client meetings. The disclosure here is light-touch — a line on the website about AI-assisted research, because the final judgement and recommendation is entirely the adviser's own.

Mid-size practice, portfolio stress-testing: Runs AI-assisted scenario modelling (market downturns, inflation shocks) as an input to portfolio reviews. This is closer to the recommendation itself, so it warrants a more specific disclosure in the SOA, plus an internal policy that every scenario output gets a human sanity-check before it's discussed with a client.

Practice using AI to draft client newsletters: Lowest-risk use case — AI drafts, a human edits and approves every word before it goes out. Still worth a line of transparency ("drafted with AI assistance, reviewed by our team") because newsletters increasingly get fact-checked by AI-savvy clients themselves.

How to build a disclosure practice that actually holds up

1. Audit every tool currently touching research, analysis or client communication — including ones staff adopted informally without telling the practice principal.

2. Separate tools by proximity to advice: drafting and admin tools need lighter disclosure than tools feeding directly into recommendations.

3. Draft specific, plain-English disclosure language (see the line-up above) rather than generic boilerplate, and run it past your licensee or compliance adviser.

4. Build in a mandatory human review step for anything AI-assisted that reaches a client, and document that review happened.

5. Revisit the policy every six months — this is a fast-moving area and today's "fine, no disclosure needed" tool can become tomorrow's "needs a clear statement" tool.

💡 Disclosure can be a differentiator, not just a compliance chore. A short, confident "here's how we use AI and here's how we don't" page tends to land as more trustworthy than silence — clients increasingly assume every practice is using some form of AI, and the ones who explain it plainly read as the most in control of their process.

Mistakes to avoid

Saying nothing and hoping it never comes up — it will, and being asked directly without a ready answer looks worse than any disclosure would.
Letting AI-generated analysis go to a client without a documented human review step.
Using generic, copy-pasted AI disclosure wording without checking it against your own licensee's requirements.
Treating AI tools as interchangeable with a paraplanner or adviser's judgement — they're not, and overselling their reliability in client materials is a credibility risk.

Frequently asked questions

Do I need to disclose AI use for every single task, even internal admin?

No — the line is proximity to the client's outcome. Internal scheduling or note-taking tools don't need client-facing disclosure; anything feeding into research, analysis or recommendations generally should be mentioned somewhere a client would see it.

Can AI tools replace a paraplanner's work?

Not reliably, and this is worth being honest about rather than oversold — AI tools are strong at summarising and speeding up first drafts, but they can misstate figures or miss context in ways a trained paraplanner would catch. Treat AI as a speed tool under human review, not a replacement for that review.

What if a client asks directly whether AI was involved in their advice?

Answer plainly and specifically — what the tool did, and that a human adviser reviewed and is accountable for the final recommendation. Vague reassurance ("don't worry about that") tends to damage trust more than an honest, specific answer would.

Is there a standard, industry-approved disclosure wording we should just copy?

Not yet a single universal standard — this is an area regulatory guidance is still catching up to, so any wording should be treated as a starting draft to run past your own licensee rather than a fixed, final answer.

Keep reading 🤍

Written by

Kate, founder of Chronically Online

I help Gold Coast and Brisbane businesses grow with branding, websites and marketing that actually works.

Work with me ✦