On the historical road to large language models (LLMs) and generative AI, there is the story of ELIZA, the first computer program to use natural language processing, developed by Joseph Weizenbaum in 1966. Weizenbaum’s secretary started conversing with the program and noted that ELIZA seemed to have human-like feelings after actively listening to her ramblings. It made Weizenbaum realize that even short interactions with a rudimentary program could illicit “powerful delusional thinking” in the average person.
Now, LLMs are so advanced that they’re built into search engine functionality, answering user questions and only showing a select few sources before the user scrolls down to the search results. Having your community financial institution’s (CFI) content make it into the AI summary is the kind of exposure that shines a light on your credibility, trustworthiness, and authority on any given topic.
What does it take for your CFI’s website content to be included in AI search engine summaries? Three experts in digital and creative marketing strategy detail what the LLM crawlers are looking for and how you can aim for that top spot in AI searches.
Expand Your Content Where typical non-AI search engine results will yield websites that answer a person’s core question, AI searches using LLM crawlers will choose sites with expanded content. The AI answers not only address the core question, but also follow-up questions the user would likely think about next, says Wayne Cichanski, VP of Search & Site Experience at iQuanti.
For example, if a person asks how APR is calculated, websites would stand a better chance of being included in AI-generated summaries if they not only answered that question, but also how an APR on a loan is different from its interest rate, as well as the difference between a fixed APR and a variable APR. Cichanski recommends creating a “mini-knowledge hub” to establish your content as a trusted, authoritative source on the subject.
These are also called “topic clusters” across multiple pages of a website, which bolsters the idea that the website provides more holistic and comprehensive content, says Manish Waskar at Thrive Internet Marketing Agency. Each page can address a subtopic and link to the main page that defines a concept, making it easier for LLM crawlers to extract content from each page.
With your topic cluster, you can build a knowledge base of neatly, clearly segmented subtopics that answer common questions your customer base might use a generative search to ask.
Structure Your Content
Structured formats like step-by-step guides, comparison guides, checklists, listicles, and FAQs are easy for LLM crawlers to extract and cite, Waskar says. “Use descriptive headers that match real questions. Avoid vague H2s like ‘Overview’ or ‘Things to know’.”
The first thought within content material should answer the user’s core question directly with precise, clear statements. Use specifics and statistics instead of making broad, general statements that can’t be quickly summarized by the LLM.
Devise shorter paragraphs that focus on one idea at a time. LLMs have an easier time summarizing this way, instead of having to parse through long, dense material.
Be an Authoritative Voice
LLM crawlers also look for content that sounds like human subject matter experts created it, according to Holinex Digital’s team. Real author names or actual case studies should be used, as well as links to reputable sources to bolster any claims.
Update older statistics and figures to reflect the most current data, because the AI will skip over content that appears to be outdated. Even updating the publish date or last-updated date of your content can help flag it as fresh.
Check Your Work
Cichanski recommends referencing a checklist like the one below before posting content on websites:
- Is the content retrievable? Is it authoritative, and can LLM crawlers easily discover it deep within webpages?
- Is the content extractable? Can LLM crawlers pull content easily from succinct statements and short paragraphs?
- Is the content citable? Does the content include quotes or stand-alone, precise, clear sentences that can be used in summaries without the need for paraphrasing?
- Is the content reusable? Can LLM crawlers use the content over and over again in subsequent searches, making the AI generative process more efficient over time?
“Make sure your content is snackable, chunkable and stackable,” Cichanski says.
The advent of AI search summaries is changing how we write and present content for our customers. While SEO is still a factor in your business’ search engine discoverability, AI-powered search result crawling has a real effect on what sources users see first when they search.
By following these strategies, you can increase your chances of being included in an AI-generated search engine summary and establishing your business as a trusted authority in that search topic.
