Ask Anything: Six Weeks of Investors Putting ETF Questions to AI

ETF.com Staff

Thu, August 20, 2026 at 9:01 PM GMT+3 4 min read

ETF Investing Tools

Since early this summer, visitors to ETF.com have been able to ask an AI assistant anything about ETFs — in plain language, in any language. Here is what more than five hundred of them did with it.

Key Numbers

568 investors asked 1,257 questions between July 7th and August 2, 2026
74% of questions were typed in users' own words rather than clicked from suggested prompts
Questions arrived in six languages, including Chinese, Hebrew, Turkish and Portuguese
Weekly volume climbed to a peak of 360 questions in the week ending August 2 — about 2.4× growth in four weeks

The Conversation Is Growing — Fast

The assistant launched quietly, with a preview at the end of June. Since the full rollout in early July, weekly questions have climbed from roughly 150 to a peak of 360 in the week ending August 2 — with investors now asking around 40 questions a day.

Bar chart: questions asked per week, growing from 130 in late June to a peak of 360 in the week ending August 2

Less "What Is an ETF?" — More "Which One Is Right for Me?"

Read a thousand of these questions and a picture emerges: this is not a glossary lookup tool. Investors treat the assistant like a research analyst. The single biggest use is screening and discovery — "show me international ETFs with over $1B in assets," "which bond ETFs have the lowest expense ratios," "list the top ten YTD-performing Capital Group ETFs." Close behind comes genuine learning: how expense ratios compound, what tracking difference means, how to read a K-1.

And then it gets personal. Dozens of investors asked the assistant to help them build or stress-test a portfolio:

"Construct a ETF portfolio for me, I retire in 20 years, aggressive risk, starting value of 1.2 million, keep it simple 3-4 funds."
— an ETF.com user, July 2026

Bar chart of question themes as share of user-written questions: screening and discovery leads at 19.4%, learning the basics 13.3%, dividends and income 7.3%, comparisons 6.5%, fees and taxes 6.1%, fund flows 6.1%, portfolio building 5.7%, hot themes 4.0%

Fund-flow questions are a quiet standout — a theme almost unique to ETF.com's data. Users asked how many days feed the daily, weekly and monthly flow reports, which funds topped each list, and "how much money has flowed into nuclear and uranium ETFs this year?"

The Tickers on Everyone's Mind

The S&P 500 heavyweights dominate, as they do everywhere — but the mix tells its own story. SCHD shows up whenever the conversation turns to dividends, SMH rides the semiconductor wave, and comparison questions like "what's the difference between QQQ and QQQM?" and "how much overlap between SMH and SPYM?" reveal investors optimizing at a level of detail that used to require a spreadsheet and an afternoon.

Bar chart of most-mentioned tickers as share of user-written questions: SPY 2.9%, QQQ 2.8%, VOO 1.9%, SCHD 1.7%, SMH 1.4%, QQQM 1.3%, ARKK 1.2%, VTI 1.1%, SPYM 1.1%, VNQ 0.9%

Beyond the majors, curiosity runs to the frontier: which ETF holds the most SpaceX, how to invest in the supply chain behind Nvidia's Vera Rubin platform, aerospace-and-defense funds that beat PPA, and whether a new muni fund is triple tax-exempt.

Suggested Prompts Open the Door — Then People Make It Their Own

A set of starter prompts gives first-time users a nudge, and one clearly strikes a nerve: "What's the cheapest way to invest in the S&P 500?" was clicked more than a hundred times, twice as often as any other. Cost is the door most investors walk through.

But the prompts are a beginning, not the product. Three out of four questions are typed in the user's own words, and over half of all questions arrive as part of a back-and-forth conversation — users pushing on an answer, narrowing a list, asking "why not VOO?" One in five users asked three or more questions.

Bar chart of starter prompt usage as share of all starter-prompt uses: the cheapest way to invest in the S&P 500 leads at 32.8%, AI exposure 16.4%, VOO vs IVV vs SPY 15.2%, QQQ vs QQQM 12.2%

Conversations That Go Somewhere

The most striking sessions read like a meeting with an advisor. A Brazilian investor spent fourteen questions working from "can you break down the US ETF market into categories?" all the way to which tab on ETF.com displays median tracking difference. Another user walked step by step through REIT ETFs — comparing USRT and SCHH, then drilling into sector allocation by property type. A third asked the assistant to review a rules-based momentum framework they were developing.

"If you were 27 yr old what would u be investing in and why?"
— an ETF.com user, getting to the point

And the audience is global: questions arrived in Chinese, Hebrew, Turkish, Portuguese and Russian alongside English — including a researcher in Chinese requesting the year-by-year count and assets of US active ETFs back to 2008.

What It Adds Up To

Six weeks in, the pattern is clear. Investors do not want another search box — they want to ask the next question. They start with cost, follow with comparison, and end up somewhere no static screener page could take them: their own portfolio, their own time horizon, their own hunch about the next big theme. The AI assistant turns ETF.com's data — flows, holdings, performance, classifications — into a conversation, and users are treating it exactly that way.

To take advantage of ETF.com's Data AI tool visit the page here - https://www.etf.com/etf-data-ai

Methodology: analysis of 1,257 questions asked by 568 unique users of the ETF.com AI assistant between June 25 and August 6, 2026. Internal test accounts excluded. Theme counts are based on questions users typed themselves (excluding suggested starter prompts); a question may count toward more than one theme. Quotes are anonymized user questions, lightly edited for capitalization and punctuation only.