Keyword research as a discipline emerged when search behaviour was largely 2-3 word queries and ranking algorithms heavily weighted exact-match keyword density. The discipline has evolved continuously as search has become more conversational, more intent-aware, and now increasingly mediated by LLMs. The 2026 best practice combines several methods. **What hasn't changed:** - Understanding what users are searching for remains foundational. - Volume estimates still inform prioritization (high-volume queries deserve more substantial content investment). - Competition assessment still matters (some queries are cost-effective to pursue; others aren't). - Intent classification (informational, navigational, transactional, commercial investigation) still drives content strategy. **What has changed:** **Topic clusters over individual keywords.** Modern SEO content rarely targets single keywords; it targets topic clusters that include the primary query plus related queries and natural variations. A page on 'small business SEO Canada' should naturally cover 'how much does SEO cost Canada,' 'how long does SEO take Canada,' 'SEO agencies Canada,' and dozens of related queries — because users searching the primary query will reach the page from many adjacent queries, and because LLMs synthesizing answers prefer pages that comprehensively cover the topic. **Conversational and long-tail patterns.** Voice search and AI chat have driven natural language patterns. 'Best family dentist accepting new patients in Barrhaven for kids under 10' is the kind of query users now ask, especially via AI engines. Keyword research that focuses only on short-tail patterns ('Barrhaven dentist') misses the substantive long-tail traffic that converts well. **AI-engine query auditing.** Beyond traditional keyword tools, modern keyword research includes asking ChatGPT, Perplexity, Claude, and Gemini the questions your customers ask. Record the synthesized answers and cited sources for each. The patterns reveal: which queries produce AI Overviews vs. classical results in Google; which competitors are cited as sources by AI engines; what gaps exist where no good source is currently cited (opportunities for your content to fill). **Intent-first prioritization.** Rather than prioritizing by volume alone, modern prioritization weights queries by business value and competitive accessibility. A query with 50 monthly searches that converts at 8% may be more valuable than a query with 2,000 monthly searches that converts at 0.3%. Build keyword lists with conversion potential and competitive intensity scored alongside volume. **Branded vs unbranded research.** Track branded search volume as a leading indicator of brand awareness. Growing branded search typically precedes growing direct traffic and increased referral activity. Unbranded research identifies the broader topic space; branded research validates that the brand is becoming a recognized entity. **Modern keyword research workflow:** **Step 1: Foundation research.** Use Ahrefs, SEMrush, or Google Keyword Planner to identify the broad topic space. Pull lists of queries containing your priority terms; lists of related questions; lists of queries your top competitors rank for; lists of queries with featured snippets or PAA boxes. **Step 2: Cluster the queries.** Group queries by topic cluster (related queries that should be addressed by the same page or set of pages). Identify the primary query for each cluster (typically the highest-volume informationally-substantive query) and the supporting queries. **Step 3: AI engine audit.** For each priority cluster, query the major AI engines with the primary query. Record: does an AI Overview appear in Google? What sources are cited by ChatGPT, Perplexity, Claude, Gemini? Are competitors cited? Is the synthesized answer accurate? Are there gaps where the synthesis is incomplete or wrong? **Step 4: Intent classification.** Classify each cluster by intent (informational, navigational, transactional, commercial investigation). Different intents require different content patterns and conversion strategies. **Step 5: Prioritization.** Score clusters by business value (likely conversion rate × likely value per conversion × potential traffic), competitive accessibility (how dominated by entrenched competitors is the cluster?), and strategic fit (does the cluster align with services you actually offer at margin?). Prioritize the high-value, accessible clusters first. **Step 6: Content brief development.** For each prioritized cluster, develop a content brief covering: primary query and supporting queries; user intent; content structure (lead with the answer for AI engine citation); named author and credentials; specific Canadian context to include; original data or analysis to differentiate from existing top results; schema markup to implement; internal links to add and receive. **Step 7: Production and publication.** Produce content against briefs; publish; submit via IndexNow or sitemap; monitor for ranking and citation outcomes. **Tools that have emerged for AI-era keyword research:** - Ahrefs and SEMrush both added AI Overview tracking in 2024–2025. - Profound, Athena HQ, Otterly.AI track AI engine citations. - AlsoAsked.com surfaces the People Also Asked patterns useful for clustering. - Google Search Console added AI Overview impression data in late 2025. Keyword research isn't dying; it's broadening. The fundamental discipline of understanding what customers search for and producing the best content to serve those searches remains foundational. The methods have multiplied, and the success metrics now include AI citation visibility alongside classical ranking position.