Empty inventory. The pattern nobody publishes.

The eight buckets.

Every Hint Co. audit splits 30 buyer questions into eight intent categories. Seven of them are what most SEO tools measure: category discovery ("best CRM"), comparison ("HubSpot vs Salesforce"), segment fit ("CRM for a 20-person team"), evaluation ("is HubSpot worth the price"), pricing, trust, switching. The eighth is different.

The eighth bucket is problem-first. The buyer describes a symptom. "Our email onboarding is not working." "Our SaaS trials are not converting." "Our reps do 50 calls a day and book zero meetings." These are the questions people ask when they do not yet know what to buy. They are the questions AI is uniquely suited to answer, because the buyer wants a diagnosis, not a comparison table.

The empty-inventory finding.

Across seven audited categories, roughly 65% of problem-first responses name no vendor at all. The answer is generic advice: a framework, a diagnostic sequence, "you might need a CRM, or a customer data platform, or maybe you have a data hygiene issue." The AI engine cannot confidently route to a specific brand because no source in its training data or search index has claimed the answer.

Why this matters.

Category-named queries are competitive. The engines have hundreds of sources for "best CRM" and they route to established players. Problem-first queries are wide open. The engines are looking for something to say, and they will cite whoever has published the specific answer to the specific symptom.

An incumbent that ignores problem-first is ceding the highest-signal buyer to whoever writes the diagnosis. Two of the last twelve audits show the category leader lost the problem-first bucket to a competitor entirely. In EX-05 (sales engagement), Apollo owns 52.2% of category-named responses. Loom, a video tool, wins problem-first at 11 mentions vs Apollo's 4. Apollo is not competing where the intent is highest.

What content wins problem-first responses.

• Diagnostic sequences. "If X, check Y; if Y is fine, look at Z." The engines cite explicit if-then structures because they can be extracted as a partial answer.
• Named symptoms with named workflows. "For a 20-person SaaS support team processing 2,000 tickets a month, automated triage reduces manual categorization but should not close billing or security tickets automatically." Specific enough to be quoted, defined enough to be useful.
• Failure stories. "Teams that try X first usually hit these three failure modes. Here is why. Here is what to check before you switch." Failure content compounds because it maps to buyer skepticism.
• Original data on causes. "In our sample of N, 43% of the cases were caused by Z rather than Y." Original data is quotable, defensible, and hard to replicate.

The fix list.

If your brand's problem-first appearance in your category audit is below 15%, the read's fix list will typically include (in this order): map the top 10 problem-first queries your buyers actually type, write one direct-answer page per query, publish the primary data behind each answer, seed the terms in credible third-party sources so the engines find the language somewhere other than your own site. Sequencing matters. See the read.

Reference audits.

Read the full audit corpus at the research library. Each case file publishes the problem-first bucket in full, including the zero-vendor rate and the verbatim excerpts of responses where no incumbent was named.