The specific considerations for investment brands in AI search
Key takeaways:
- Financial advisors and investors are gravitating to AI search.
- What helps in AI search results: Attention that investment firms have gained across the web, thanks to financial PR, content marketing and social efforts.
- To asset managers’ disadvantage, LLMs may struggle to 'read' fund profiles.
If you're like many digital marketers in 2025, you may be consumed with three questions related to the growing reliance on large language models (LLMs) such as ChatGPT, Perplexity, Claude, Gemini, Microsoft Copilot and others for internet searches:
- To what extent are my clients and prospects gravitating to AI search?
- How do my brand and products fare in searches that are taking place?
- What are the implications for my website and business strategies that rely on engagement on my domain?
By now these are well trodden topics on digital marketers’ favorite sites. But there are special considerations when attempting to get a grip of questions related to investment company brands, financial advisors, and other professional and retail investors.
Settle in while we take a crack at addressing all three, as informed by our continuing work on the Lowe Group Visibility Grader. This discussion is particularly germane for firms in motion. If you’re about to update your website or assimilate another firm and its products, you’ll want to pay special attention.
Advisor and investor use of AI search
Advisors and high net worth investors tend to be early adopters, and data published in the last few months confirm a growing reliance on search beyond Google (and let's not forget TikTok).
More than 40% of advisors use AI search for business, we learned from the T3/Inside Information survey published last month. “We doubt it will be long before this category’s market penetration figure rivals CRM and Financial Planning,” according to the report.
Affluent investors, too, “are embracing artificial intelligence (AI)-powered solutions to summarize market insights, identify new investment opportunities, and guide their allocations,” according to the Escalent Investor Brand Builder findings published in January. Almost one-third (30%) are leveraging AI to inform their investment decision-making. Of those, nearly half (46%) say they use the tools daily.
Organic vs. AI searches
Let’s first acknowledge the difference between success in Google search and AI search because it points to an inherent advantage for investment company brands.
As SparkToro’s Rand Fishkin has said:
The currency of Google search was links. The way that you ranked in these results was through links, relevant content, smart keyword use, and references to your work from sources the search engines crawled.
The way that you rank in large language models is not that. The currency of large language models is not links. The currency of large language models is mentions (specifically, words that appear frequently near other words) across the training data.
And there’s the advantage—the investment industry’s collective investment in public relations, including media relations, content marketing, including content syndication, and social media has assured the brands’ appearance across the web. Among advisory firms, participation in awards and testimonials/reviews achieves the same goal. Given its documented effectiveness over time, the importance of being in the conversation on others’ sites will continue to heighten the business case for financial PR.
Too, consider this advantage that not all products in all industries have: the multiple media and brokerage websites (Yahoo Finance, Morningstar, Fidelity et al) that feature detailed information about investment products.
These two advantages assure that the brand and products are mentioned well beyond their own domains, working themselves into the data that is then synthesized and used in responses to AI search queries.
“The currency of Google search was links... The currency of large language models is mentions.”
Rand Fishkin, SparkToro
Questions to ask
How does your firm fare, in particular? If you haven’t yet, we recommend you submit some random queries to the AI search engines.
Start with questions about the brand itself and you should see links to your own communications among the returned citations:
- How much do you know about Firm XYZ?
- What’s Firm XYZ known for?
- Name some of the notable people associated with Firm XYZ.
Next ask questions designed to reveal the models’ understanding of your firm in context. Try:
- Which investment firms are prominent on this subject __________?
- Which investment firms are the most searched for in the ______ category?
- Which firms are market share leaders in this ___________ category?
These are the kinds of questions a reporter might ask when doing preliminary research. Is your firm being surfaced where you’d expect? You might be surprised, both by the responses returned and by the website content being cited.
Mutual fund, ETF searches
As we’ve said, your site needs to rank first in Google search for your ticker symbols. For the foreseeable future, human searchers are likely to continue to prefer to use traditional search as the quickest way to navigate to information about a specific fund. Your product page needs to be the first, not even the second, result.
Similarly, you’ll want to be in the mix as non-navigational product-related searches shift to AI. As the manufacturer/issuer of your products, your firm certainly has the most to offer to searchers, including advisors and the media.
In reviewing the visibility of your site among AI responses returned and citations provided, your goal must be for the product information available on your site to be accurately represented. And yet, product searches are where we see the greatest gaps between what the models are finding and what we know to be true.
In my targeted querying about funds, “wrong” answers surface in a few ways, including omissions and mis-statements. One maddening comment that appears consistently about mutual funds is that the LLMs consistently seize on the common $1 million minimum investment for Class I shares, declaring them inappropriate for most retail investors. What’s understood offline in the intermediary channel needs to be documented on a web page that the models can train on.
Beyond that, I’m seeing publishing and presentation issues across the board. It looks as if the issues are most prevalent for firms whose website presentation of fund data varies from fund to fund and those that rely on unstructured data display. You’ll be at a particular disadvantage unless adjustments are made.
“The LLMs consistently seize on the common $1 million minimum investment for Class I shares, declaring the funds inappropriate for most retail investors.”
Visibility assumes readability
My guess is that none of what follows will surprise IT experts, but the importance of presentation and structure of your most important content (product pages) has now come home to roost for those responsible for raising and maintaining brand and product visibility. As mentioned previously about marketing emails, you’re publishing for machines now—smart machines, to be sure—and what you publish needs to be machine-readable.
Developing an awareness of what’s needed to be among the sources cited will make you a better buyer and internal or external client of those you rely on to develop and maintain your site.
I test these searches, looking for the:
- Largest funds in a category—i.e., to confirm that my target fund’s AUM is readable.
- Top performing—confirming the performance data tables
- 5 star funds—can they find the Morningstar stars and category information, and are the rankings readable?
- Cheapest—confirming the expense data (which can be the source of some stumbling because of asterisked information)
- Capital gains distributions paid
- Portfolio managers associated with funds
Here’s a sample of the back and forth:
Lowe Group: List the five star-rated funds in Morningstar category X.
Search returns a response that includes a list of funds, but it’s missing the fund I’m looking for.
Lowe Group: Fund name (ticker symbol) has five stars in Morningstar category X, why didn’t you include it in your list?
“Ticker symbol X does not have a five-star rating.”
OK, timeout, what would a typical searcher likely do at this point? If they’re a reporter looking for some background on a fund’s competitive set, they might double-check their notes from a conversation with the fund company’s PR person. An advisor might return to a wholesaler’s email. An investor might reach for a fact sheet and squint to take a closer look at the disclosure. The point is that confidence has been undermined and there’s a good chance that the searcher will conclude that the AI list is correct and that their information is either wrong or out of date.
But when I’m searching, I’m not giving up. I follow up with proof—I actually upload a screenshot from the fund profile itself. And, the capitulation by AI is wholly unsatisfying.
Lowe Group: Do you see the five stars in the screenshot from the fund profile?
You are absolutely right! My apologies for the oversight. AI is still learning.
How realistic is the search above, especially when professionals have so many other ways and tools for vetting funds?
The five-star search, worded as such, may never be submitted again but it’s not about those words per se. In the good old days of two- and three-word Google searches, it was easier (and effective, too) to optimize based on common or known searches as reported in SEO tools. A Google search for “top performing large cap growth funds” will return results with links to pages that have those very words, as you can see.
Generative engine optimization (GEO) is different. The value of conducting these test searches is to reveal context. An AI searcher might ask for “top performing funds” in a Morningstar category and these engines are smart enough to look for five-star funds. Not appearing in the five-star search result reveals that the fund will likely not be included in any searches related to top performing funds.
“An AI searcher might ask for ‘top performing funds’ and these engines are smart enough to look for five-star funds.”
Technical limitations that impede results
So why are some funds being overlooked?
Below I’ll share what I’ve been able to piece together, repeatedly asking the models to distinguish between best and troublesome publishing approaches.
Investment product data that LLMs struggle with
- Sites that rely heavily on client-side rendering through JavaScript may not present their complete information when accessed through some search and crawling methods. In those cases, only the template structure is captured—{{ }} formatting—instead of actual values rather than the fully rendered content that would be visible to a human user accessing the page directly through a web browser. LLMs prefer server side rendering that generates the complete HTML on each request before sending it to the client.
Here's Perplexity’s comment (anonymized) about a fund profile it couldn’t read:“I cannot definitively determine whether ABCD has a five-star Morningstar rating or its current assets under management. The technical limitations in how the webpage content was captured and processed have resulted in template placeholders rather than actual data being available for analysis. This rendering issue explains why ABCD may have been omitted from a previous list of five-star funds, as the actual rating could not be verified from the available information.
- Content organized in a hybrid format combining tabular data with descriptive text sections. More sophisticated extraction tools—requiring more time, in other words—may be needed to properly distinguish between quantitative data and qualitative explanations.
- Multiple locations for key data. Example: AUM is presented in different formats and locations (e.g., table vs. strategy overview), which can lead to confusion or misinterpretation.
- Contextual clarity. Example: AUM that lacks context about whether it refers to a specific share class or the entire fund.
- In order for the search engines to see your stars, you need to include them on your fund profiles. It’s not enough to show them in the fund’s fact sheet PDF. “This requires extra steps to access, which may not be optimal for quick recognition or automated processing,” according to what the models told me.
Investment product data the LLMs can read
- HTML tables that create a clear structural hierarchy that separates different data categories and maintains relationships between related values.
- Performance data that follows an identical structural pattern, enabling straightforward algorithmic comparison across periods.
- Portfolio composition data organized into discrete tables with consistent formatting. Strategy allocation percentages, duration metrics, credit quality breakdowns, and maturity distributions all appear in dedicated table structures with clear column headers and value assignments. A consistent approach to data categorization creates a predictable pattern that automated tools can reliably parse.
- Numerical data that exhibits strong formatting consistency, which significantly improves machine processing capability.
- Consistency from fund profile to fund profile.
Tip: When you do your own test searches, using Deep Research mode will show you the series of steps and “reasoning” the model is applying to produce a response. It’s fascinating.
What more AI searches mean for your site
Asset manager websites already go head to head in Google with prominent, well resourced websites. Acquiring organic search traffic for more than brand terms has always been a challenge requiring deliberate focus.
Even so and notwithstanding the growing reliance on AI search, you likely have an ongoing reason to want to bring as many visitors as possible to your site. Your digital engagement strategies likely depend upon it.
What are your prospects? I’ve been talking to the LLMs about that, too. Is it possible, I’ve wondered, to isolate factors that would help the search engines recognize your authority and show your site preference in the citations offered along with their synthesized responses to searches?
While not at all groundbreaking, their comments provided insight and possibly inspiration for you and your team to continue to work on optimization.
“From the perspective of an AI search engine like myself, determining whether to rely on the fund company's profile page or a media website depends on how the data is structured and presented for ease of processing.”
The advantages of a media website, according to the models, include:
- Media websites like Morningstar or Bloomberg often use structured, tabular formats that are AI-friendly.
- Key metrics are labeled clearly and presented in a standardized manner, making it easier to extract, compare, and rank funds.
- The inclusion of third-party ratings (e.g., Morningstar stars) adds credibility and context that fund company pages may not always highlight.
Fund company limitations include:
- While fund company websites provide authoritative data, they often embed key metrics within narrative text or less-structured sections (e.g., "Strategy Overview").
- Discrepancies in figures (e.g., AUM differences) can arise if the website does not clarify whether values apply to specific share classes or total assets.
- The absence of structured data markup (e.g., schema.org) makes it harder for search engines to extract information directly from the page.
How to track AI-sourced traffic
- Create a GA4 Free form exploration report
- Dimensions: Date, Page title, Session/source/medium and Metrics: Sessions, Views, Average session duration
- Filter by choosing Matches regex and paste this regex pattern to target the predominant domains:
^https://(www.meta.ai|www.perplexity.ai|chat.openai.com|claude.ai|chat.mistral.ai|gemini.google.com|chatgpt.com|copilot.microsoft.com|copy.ai)(/.*)?$|.*.ai.*|.*.openai.*|.*.groq.*|.*.metaai.*|.*.meta.com/ai.*
If you haven’t focused on your firm’s showing in AI search yet, we encourage you to—and we’re ready to help. To explore the possibilities, send us a note.
Subscribe.
Receive the latest news and insights from Lowe Group.