Methodology Memo
Subject: Substantiation methodology for "first AI-managed" ETF claim
Scope: U.S.-registered ETFs
Search dates: February 2, 2026 through February 4, 2026
Purpose: To document the methodology used by FINQ to substantiate the claim that its ETFs are "first AI-managed" ETFs, for FINQ's internal substantiation file.
METHODOLOGY
FINQ reviewed the U.S.-registered ETF universe using the following sources:
- SEC EDGAR
- Bloomberg
- ETF.com
- ETFDB.com
- ChatGPT
- Gemini
- Grok
- Claude
- SEC EDGAR was treated as the primary source for official ETF disclosure language, including prospectuses, statements of additional information, registration statements, and supplements.
- Bloomberg, ETF.com, and ETFDB.com were used as secondary sources to identify potentially comparable ETFs and cross-check fund names, tickers, categories, and descriptions.
- ChatGPT, Gemini, Grok, and Claude were used only as research tools to help identify potentially relevant ETFs and search terms. FINQ did not treat AI-tool outputs as primary substantiation. Any potentially relevant ETF identified through an AI tool was subject to further review using SEC EDGAR, Bloomberg, ETF.com, or ETFDB.com.
SEARCH TERMS
FINQ used the following search terms:
- AI
- artificial intelligence
- machine learning
- ML
- deep learning
- large language model
- generative AI
- GenAI
- NLP
REVIEW PROCESS
For each potentially comparable ETF, FINQ reviewed the fund name, ticker, adviser/sub-adviser, listing exchange, prospectus strategy language, SAI language where relevant, and whether the AI-related claim related to the fund's holdings theme or its actual portfolio management process.
FINQ focused on distinguishing ETFs that invest in artificial intelligence-related companies from ETFs whose investment selection or portfolio management process is itself driven by artificial intelligence.
RELEVANT RESULTS IDENTIFIED
FINQ identified potentially comparable ETFs that use AI-related terminology or AI-related investment processes. The results table retained in the substantiation file includes, among others:
- WisdomTree AI Enhanced Value Family — AIVI/AIVL;
- QRAFT AI-Enhanced/Powered ETFs — QRFT/AMOM/LQAI;
- Stocksnips AI-Powered Sentiment US All Cap ETF — NEWZ;
- Amplify AI Powered Equity ETF — AIEQ.
FINQ's review found that these identified ETFs involved one or more distinguishing features, including human portfolio manager review or adjustment of AI model output, adviser discretion over investment decisions, portfolio manager selection from an AI-generated list, or passive index tracking. The results table documents the relevant prospectus language, fund strategy summaries, and use of AI for each ETF reviewed.
ADDITIONAL ETFs REVIEWED AND DETERMINED NOT COMPARABLE
| ETF | Ticker | Why it appeared in search | Why excluded from main comparable table |
|---|---|---|---|
| ETFVanEck Social Sentiment ETF | TickerBUZZ | Why it appeared in searchUses AI-related index terminology, including the BUZZ NextGen AI US Sentiment Leaders Index | Why excluded from main comparable tablePassive/index-tracking ETF; AI-related element appears tied to index methodology/social sentiment analytics, not an AI-managed ETF portfolio process |
| ETFThemes Generative Artificial Intelligence ETF | TickerWISE | Why it appeared in searchFund name and strategy relate to generative AI | Why excluded from main comparable tableAI-themed/passive index ETF tracking companies with AI-related business operations |
| ETFREX AI Equity Premium Income ETF | TickerAIPI | Why it appeared in searchFund name includes "AI" and strategy references AI-related companies | Why excluded from main comparable tableAI-sector/index exposure plus covered-call income strategy; not an AI-managed stock-selection ETF |
| ETFiShares U.S. Equity Factor Rotation Active ETF | TickerDYNF | Why it appeared in searchDisclosure references machine learning and artificial intelligence methods | Why excluded from main comparable tableActive factor-rotation ETF using a broader factor model; AI/ML appears as part of a broader data-driven process, not as an AI-managed ETF strategy |
FINQ retained these prospectuses in the substantiation file to document that these search results were reviewed and excluded from the main comparable analysis.
Comparable AI-managed ETF results
| ETF Name | Symbol | Human intervention | Prospectus Quote | Fund Strategy | Use of AI |
|---|---|---|---|---|---|
| ETF NameWisdomTree AI Enhanced Value Family | SymbolAIVI/AIVL | Human interventionThe portfolio manager has the authority to adjust the AI model output before implementing it on the ETF | Prospectus Quote"The virtual portfolio manager enforces various constraints and portfolio optimization methods on the securities, while the human portfolio manager reviews the resulting portfolio and adjusts as necessary before implementing in the fund." (page 3) | Fund Strategy"The WisdomTree AI Enhanced Value Fund Family seeks to offer uncorrelated returns streams from the value universe by leveraging the expertise of the Voya Equity Machine Intelligence (EMI) team and their fundamentally driven machine learning approach." The EMI model "approaches value investing dynamically to avoid narrow style biases that may be out of favor, while still providing value exposure across equities." The strategy "seeks to capitalize on short- and long-term investment opportunities – while delivering the virtue of patience and agility, acting quickly and decisively when opportunities arise." The process "consists of five stages with varying degrees of involvement between human and machine – something the EMI team calls 'human in the loop'," emphasizing "that the strategy not only is driven by AI, but also has the proper human intervention and oversight to ensure robust security selection and risk management." | Use of AI"The first step in the process consists of data aggregation consisting of 10,000+ data points for each company over a 20-year history." These data points are "feature engineered by human experts to provide a more insightful view of the company versus its peers, the entire stock universe, or its own historical characteristics." "Trained on 20 years of these historical features, the virtual analysts (the true AI) take the most recent data points as input to identify companies for inclusion in the portfolio." The "virtual traders use machine learning to identify dynamic 'rules' and patterns for defining value at the company level," and "focus on shorter-term indicators, identifying proper entry and exit timing, as well as any risk events." Once approved, "these security selections are passed on to the virtual and human portfolio managers for review before being executed by the human portfolio management and trading team at Voya." Ultimately, "the model continually ingests new information and tracks the features representing the changing market conditions and dynamics of the company," seeking "to capitalize on asymmetric risk/reward patterns." |
| ETF NameQRAFT AI-Enhanced/Powered | SymbolQRFT/AMOM/LQAI | Human interventionThe portfolio manager has full decision-making power | Prospectus Quote"The Adviser has full discretion over investment decisions for the Fund. Therefore, the Adviser has full decision-making power not only if it identifies a potential technical issue or error with the U.S. Large Cap Database, but also if it believes that the recommended portfolio does not further the Fund's investment objective" (page 1,4) | Fund Strategy"The Fund is an actively-managed exchange-traded fund ("ETF") that seeks to achieve its investment objective by utilizing an investment strategy enhanced by the use of artificial intelligence." Under normal circumstances, "the Fund invests at least 80% of its net assets… in securities of U.S.-listed large capitalization companies," including "common stock, American Depositary Receipts ("ADRs"), and Global Depositary Receipts ("GDRs")." The Fund's adviser, "Exchange Traded Concepts, LLC (the "Adviser"), uses an investment process based on a proprietary artificial intelligence security selection process that extracts patterns from analyzing data, developed by QRAFT Technologies, Inc. ("Qraft")." The Fund "expects to hold 300 to 350 companies in its portfolio," and while the Adviser generally follows Qraft's recommendations, "the Adviser has full discretion over investment decisions for the Fund," including adjustments for "corporate actions, mergers and spin-offs." | Use of AI"In pursuing the Fund's investment objective, the Adviser consults a database generated by Qraft's AI Quantitative Investment System ("QRAFT AI"), which automatically evaluates and filters data according to parameters supporting a particular investment thesis." QRAFT AI "selects and weights portfolios of companies… to provide a balanced exposure to a variety of factors… including quality, size, value, momentum, and volatility." Using "deep learning technologies" and "Bayesian neural networks that estimate the uncertainty of its forecast," QRAFT AI "estimates each stock's relative superiority of price appreciation… for the next four week investment period" and "selects the top 300 to 350 stocks." It then "evaluates how each individual factor would change and/or affect a company over time, identifying the companies that have the greatest potential to outperform their U.S. large cap peers." "QRAFT AI repeats such processes every four weeks," providing updated recommendations from which "the Adviser makes or changes investments in the Fund based on the newly generated information." |
| ETF NameStocksnips AI-Powered Sentiment US All Cap ETF | SymbolNEWZ | Human interventionThe AI model recommends a list of securities, <strong>the portfolio manager chooses within that list</strong> | Prospectus Quote"The Sub-Adviser expects that under normal market conditions, up to 95% of the portfolio assets will be invested in the securities recommended by the algorithms." (page 4) | Fund Strategy"The Fund is an actively-managed exchange-traded fund ("ETF") that seeks to achieve its investment objective by utilizing an investment strategy that leverages artificial intelligence ("AI") and natural language processing to derive a proprietary News Media Sentiment Signal (the "Sentiment Signal")." The Sub-Adviser "uses an investment process based on a proprietary ranking and selection process, which is designed to identify stocks with the most positive news coverage." Under normal circumstances, "the Fund invests at least 80% of its net assets in securities of U.S.-listed large, mid and small capitalization companies," typically resulting in "a universe of approximately 350 securities." From this universe, "the algorithms rank the securities… based on the amount of positive news available about a company and its sentiment momentum," and "the top 30 to 50 stocks are included in the Fund and are equal weighted." | Use of AI"The Sentiment Signal and the Sentiment Momentum Signal are each derived using natural language processing and machine learning algorithms that transform unstructured textual data from News Sources into quantified real-time sentiment signals." In this process, "machine learning is used to classify 'financial oriented' sentences (news snippets) attributed to a company as being positive, negative or uncertain." The Sub-Adviser "believes that quantification of news sentiment can be a proxy for investor sentiment" and "employs quantitative modeling, which leverages the Sentiment Signal and a trending Sentiment Momentum Signal." To generate these signals, "StockSnips gathers news information on a daily basis and evaluates such information along with a proprietary historical sentiment dataset covering approximately 5,000 US equities." |
| ETF NameAmplify AI Powered Equity ETF | SymbolAIEQ | No adviser discretion—but via index tracking. The fund passively tracks the AI Powered Equity Index built from EquBot's AI model; the adviser rebalances to the index and "does not try to 'beat' the Index." Security selection/weights are set by the Index Provider (EquBot/Solactive), not by the adviser. | |||