What a query lattice looks like in practice: a CRM software example

The previous article in this series introduced the query lattice — a content architecture built around the specific intermediate queries that AI systems generate when handling conversational prompts, rather than around the broad topic clusters that traditional SEO and first-generation AEO favour. The concept is straightforward in theory. In practice, it is easier to understand with a worked example. CRM software is a useful category because it has a well-defined competitive landscape, a clear set of user intents, and a vocabulary gap between how buyers talk about the problem and how AI systems retrieve information about it.

<h2>Start with the user prompt, not the intermediate query</h2>

A typical user prompt in the CRM software category might be: "What is the best CRM for a 20-person sales team that uses Slack and needs Salesforce integration?" That is a conversational prompt. It is not what ChatGPT Search sends to Bing. Before any retrieval happens, ChatGPT rewrites that prompt into a set of intermediate queries. Based on the structure of the prompt, those intermediate queries might look something like: "CRM software Salesforce integration comparison 2026", "CRM tools Slack native integration small sales team", "best CRM 20-person sales team mid-market", and "Salesforce-compatible CRM alternatives pricing". Each of these is a distinct retrieval event. Each targets different content. A brand that appears in the results for all four is far more likely to be cited in the final answer than a brand that appears for only one.

<h2>Map the intermediate query vocabulary</h2>

The first step in building a query lattice for a CRM brand is to map the intermediate query vocabulary for the category. This means identifying the specific terms, qualifiers, and combinations that AI systems use when retrieving CRM information — not the terms buyers use in conversation, but the terms that appear in the intermediate queries. For CRM software, that vocabulary includes: integration names (Salesforce, HubSpot, Slack, Outlook, Zapier), use-case qualifiers (sales team, customer success, B2B, SaaS, enterprise, SMB, startup), comparison framing (alternatives, vs, comparison, pricing, features), and specificity qualifiers (2026, mid-market, 20-person, 50-seat). A brand that has content covering the combinations of these terms that appear most frequently in intermediate queries has a structural advantage at the retrieval layer.

<h2>Build the lattice nodes</h2>

A query lattice for a CRM brand consists of a set of specific answer pages — lattice nodes — each targeting a distinct intermediate query or cluster of closely related queries. For the example above, the lattice nodes might include: a dedicated integration page for each major tool (Salesforce, Slack, Outlook) that answers the specific question "does [Brand] integrate with [Tool]?" with a direct, extractable answer in the first paragraph; a comparison page for each major competitor that answers "[Brand] vs [Competitor]: which is better for [use case]?" with specific, factual differences; a use-case page for each primary buyer segment (sales team, customer success, startup) that answers the specific question "is [Brand] right for [segment]?" with concrete evidence; and a pricing page that answers "how much does [Brand] cost for [team size]?" with specific figures rather than "contact us for pricing". These are not blog posts. They are answer pages: short, specific, structured for extraction, and updated regularly.

<h2>Connect the nodes</h2>

The lattice structure comes from the connections between nodes, not just the nodes themselves. An AI retrieving the Salesforce integration page should be able to navigate to the comparison page (which references Salesforce as a comparison point), the pricing page (which includes Salesforce integration in the pricing tier), and the sales team use-case page (which mentions Salesforce as a common stack component). These connections serve two purposes: they help AI systems build a coherent picture of the brand across multiple retrieval events, and they signal to the AI that the brand has depth of coverage on the topic rather than a single isolated answer.

<h2>The vocabulary gap in practice</h2>

Most CRM software brands have content that covers the broad topic of CRM. Very few have content that covers the specific intermediate query vocabulary at the combination level. A brand might have a page about Salesforce integration and a page about pricing, but not a page that answers "Salesforce-compatible CRM alternatives pricing" as a combined query. That gap is where visibility is lost. The intermediate query is not asking about Salesforce integration or pricing separately — it is asking about the intersection. A lattice node that addresses that intersection directly will outperform two separate pages that each address only half of the query at the retrieval layer.

<h2>What this looks like in the content audit</h2>

Running the audit described in the previous article — testing representative user prompts in ChatGPT Search and recording which sources are cited — typically reveals a pattern for CRM brands. The sources that appear consistently are not the brand's main product pages or blog posts. They are comparison sites (G2, Capterra, GetApp), integration documentation pages, and specific use-case guides from specialist publications. The intermediate query vocabulary those sources are optimised for is exactly the vocabulary described above: integration names, use-case qualifiers, comparison framing, and specificity qualifiers. A brand that builds lattice nodes using that vocabulary, with content that is as specific and extractable as the comparison sites, can compete directly at the retrieval layer rather than relying on third-party platforms to represent it.

<h2>AEO Updates Takeaway</h2>

A query lattice is not a content calendar. It is a map of the specific questions an AI would need to answer in order to confidently recommend your brand. For a CRM software brand, that map includes integration pages, comparison pages, use-case pages, and pricing pages — each structured as a direct answer to a specific intermediate query, connected to adjacent nodes so that an AI retrieving one can navigate to the others. The brands that are already winning at the intermediate query layer in the CRM category are mostly comparison sites and integration directories, not the CRM brands themselves. That is a gap. Closing it requires building content that is specific enough to match intermediate query vocabulary, structured enough to be extracted, and connected enough to signal depth of coverage across the full query lattice.