Researching Products, Funds, and Strategies with AI

A client asks about a structured note their brother-in-law mentioned. A wholesaler drops off a 60-page prospectus for a new buffered ETF. You're considering whether direct indexing makes sense for a particular household. In each case you need to get oriented fast, then verify carefully. AI is excellent at the first part — and dangerous if you skip the second.

What You'll Learn

• Using AI to get oriented quickly on unfamiliar products and strategies
• Summarizing and interrogating long documents like prospectuses and contracts
• Comparing options across the dimensions that actually matter
• Why "verify against the source" is non-negotiable here

Getting Oriented Fast

When something unfamiliar lands on your desk, the first job is a mental map: what is this, how does it work, what's it for, what are the catches. AI is great at producing that map.

You are a financial advisor's research assistant. Give me a clear, neutral briefing on [buffered/defined-outcome ETFs]: what they are, the basic mechanics, the typical use case, the main trade-offs and risks, the fee considerations, and the kinds of clients they tend to fit or not fit. Be balanced — include the criticisms. About 400 words, plain language. Flag anything I should verify against primary sources before relying on it.

That briefing isn't your due diligence — it's your starting point so the due diligence goes faster. For anything where currency matters (a rule change, a recent fund event, regulatory commentary), switch to Perplexity so you get cited sources you can open and check, rather than a confident summary you can't trace.

Summarizing and Interrogating Long Documents

This is where AI earns its keep. Drop a long document into Claude (large context window) or ChatGPT with file upload — after redacting any client identifiers or account numbers, and using only your firm-approved tool if the document is confidential — and put it to work.

Here is the prospectus for [fund]. Summarize it for me as an advisor: investment objective and strategy in plain English, key risks, total expense ratio and any other fees, liquidity and redemption terms, distribution policy, and anything unusual or buried that I should pay attention to. Then list the five questions I should be able to answer about this fund before recommending it, and tell me where in the document the answers are.

Follow-ups make it a conversation:

• "What does it say about leverage or derivatives use?"
• "Compare the fee structure here to a plain-vanilla index fund in the same category."
• "Is there a surrender schedule? Lay it out year by year." (for an annuity contract)
• "Summarize the risk factors a conservative retiree would care about most."
• "Quote the exact language about how the cap or buffer is set and when it resets."

The same approach works for variable annuity contracts, REIT offering documents, 401(k) plan documents you're reviewing for a participant, trust documents that affect a plan, and SMA disclosure brochures. You get to the substance in minutes instead of an hour of careful page-turning — but you still confirm the critical terms by reading the actual relevant pages, because a summary can miss or soften something that matters.

Comparing Options

Clients and prospects constantly ask "which is better, A or B?" AI is good at structuring the comparison so you don't forget a dimension.

Build a comparison table for a client deciding between [Option A] and [Option B] — for example, a target-date fund versus a managed model portfolio. Columns: how it works, cost, tax efficiency, flexibility/customization, who it tends to suit, main downsides. Keep each cell to a phrase. Below the table, add three plain-English sentences on how I'd help a client think about the choice, without recommending one in the abstract since it depends on the individual.

Other comparisons advisors run often: traditional vs. Roth contributions, lump-sum pension vs. annuity, term vs. permanent life insurance, taxable brokerage vs. tax-deferred for a given goal, direct indexing vs. index ETFs, two similar funds in the same category. The table is a thinking aid you adapt to the real client — not a recommendation engine.

The Non-Negotiable: Verify

This lesson comes with the strongest warning in the course, because product research is exactly where hallucination does the most damage. An LLM will state an expense ratio, a surrender period, a contribution limit, a tax treatment, or a fund's holdings with total confidence — and sometimes be wrong, out of date, or describing a different product. Operate accordingly:

• Numbers come from primary sources. Expense ratios, fees, caps, buffers, surrender schedules, contribution and income limits, RMD ages, tax brackets — verify in the prospectus, the contract, the fund company's site, or the IRS, never from the AI's recall.
• Confirm which product you're actually discussing. Tickers and names are easy for a model to confuse. Make sure the document you uploaded is the one you're asking about.
• Use Perplexity or direct sources for anything current. Rules and products change; a model's training has a cutoff.
• Don't quote the AI to a client. "According to my research..." should mean your research, validated, not a chatbot's paragraph.
• Document where it matters. If a recommendation rests on a product's terms, your file should reflect that you checked the source — because an examiner won't accept "the AI told me."

Used this way, AI compresses the boring orientation-and-summarizing part of research and leaves you more time for the judgment part — which was always the part that mattered.

Key Takeaways

• Use AI (and Perplexity for cited, current info) to get a fast, balanced briefing on unfamiliar products and strategies — as a starting point for due diligence, not a substitute for it.
• Drop long documents (prospectuses, annuity contracts, plan documents) — redacted, in a firm-approved tool — into Claude or ChatGPT and interrogate them: objective, risks, fees, liquidity, surrender schedules, buried terms.
• Have AI structure option comparisons so you cover every relevant dimension, then adapt the table to the real client rather than treating it as a recommendation.
• Verify everything that matters against primary sources — fees, caps, surrender schedules, limits, tax treatment — confirm you're discussing the right product, never quote the AI to a client, and document your real checking.