Keyword research used to be simple: find a phrase with high search volume, write a page around it, and watch the traffic roll in. That playbook is breaking down. AI Overviews now answer many queries directly inside the search results page, and it’s entirely possible to rank in position one for a keyword and still get almost no clicks.
That shift isn’t something to panic about. It’s a reason to get more deliberate about which keywords you chase in the first place. This guide walks through a modern keyword research process built around buyer intent, not just search volume, along with how the process has changed now that AI shapes so much of how people search.
What is keyword research?
Keyword research is the process of identifying and analyzing the words and phrases people use to find information, whether they’re typing into a traditional search engine, asking a chatbot, or searching within a social platform.
For a long time, “keyword research” meant “Google research.” That’s no longer accurate. Search now happens across multiple surfaces, and people use noticeably different phrasing depending on where they’re searching. Someone typing into Google might search “best ergonomic desk,” while the same person asking an AI assistant might phrase it as a full question: “what’s a good ergonomic desk for someone with lower back pain who works from home.”
Google still commands the largest share of search activity by a wide margin, but a meaningful and growing slice of search now happens on social platforms and through AI chat interfaces. That’s exactly why keyword research today has to account for more than one destination.
Why keyword research still matters
It reveals what your buyers actually think and need
Search queries are one of the most honest data sources available to marketers. People type exactly what’s on their mind, in their own words, with no filter. Mining that data tells you what your audience struggles with, what language they use to describe their problems, and how close they are to being ready to buy.
It future-proofs your content strategy
Older SEO strategies chased traffic for its own sake: rank for a high-volume informational term, celebrate the traffic spike, move on. The problem is that this approach never guaranteed revenue, and now that AI Overviews answer many informational queries directly in the results page, pure traffic-chasing keywords are producing far fewer clicks than they used to.
Take a purely informational query like “how to press flowers.” A search like that is now often fully answered inside an AI summary, with little reason for the searcher to click through anywhere. Compare that to a transactional query like “buy flowers online,” which still reliably sends the searcher straight to business websites, because the intent (make a purchase) can’t be satisfied by a summary alone.
The lesson: prioritize keywords tied to an action, not just curiosity. Informational content still has a role in building topical authority and staying present in AI-driven answers, but it shouldn’t be the centerpiece of a strategy aimed at revenue.
It tells you whether you can realistically rank
Every keyword has a competitive landscape attached to it. Researching who currently ranks for a term tells you whether it’s realistic for your site to compete, or whether you’d be better served chasing a narrower, less contested variation of that topic.
A simple 3-step starting process
If you’re new to this, here’s a lightweight way to get your first keyword list together.
Step 1: Start with a seed keyword. Pick one word or short phrase that describes your business or the topic at hand, and drop it into a free tool like Google Keyword Planner or Ubersuggest.
Step 2: Check the core metrics. Look at search volume (roughly how many people search the term monthly), keyword difficulty (how hard it will be to rank), and search intent (what the searcher actually wants). You’re looking for a workable middle ground: enough volume to matter, but not so competitive that ranking is unrealistic for your site’s current authority.
Step 3: Organize what you find. Export your list, group terms by topic and by intent, and use that grouping as the backbone of your content plan. Resist the urge to chase every keyword you find; a focused list of ten to twenty well-chosen terms beats a sprawling list of hundreds you’ll never fully act on.
Understanding the core metrics
Search volume shows the average number of monthly searches for a term. More isn’t automatically better. A term with 500 monthly searches that closely matches your offering can be far more valuable than one with 10,000 searches that’s only loosely related to what you sell.
Keyword difficulty is typically scored 0 to 100 and estimates how hard a term will be to rank for, based largely on the authority of sites currently occupying the top results. Newer or smaller sites generally have the best luck in the lower difficulty ranges, while highly competitive terms tend to be dominated by large, established brands.
Search intent explains the “why” behind a search. The four commonly recognized categories are:
- Informational — the searcher wants to learn something (“what is a CRM”)
- Navigational — the searcher is trying to reach a specific site or brand (“Salesforce login”)
- Commercial — the searcher is comparing options before deciding (“best CRM for small teams”)
- Transactional — the searcher is ready to act (“buy CRM software”)
Getting intent wrong is one of the costliest keyword research mistakes. A perfectly optimized page will still underperform if it doesn’t match what the searcher actually wants at that moment.
Building a keyword strategy step by step
1. Start with topics, not keywords
Before diving into individual terms, define one to three broad topics you want your site to be known for. Trying to cover too many topics at once spreads content resources thin without building real authority in any of them. It’s usually better to build depth in a narrow set of topics first, then expand once those areas are performing well.
2. Identify the keyword behind your conversion page
Work backward from the outcome you want. If you’re selling a product or service, find the commercial and transactional keywords that map most directly to that offering, and make sure you have a page built specifically to convert people who search those terms. That page should make it obvious what you offer, what it costs, and how to take the next step.
A quick way to sanity-check intent without paid tools: search the term yourself and see what Google is already showing. If the results page is full of product or pricing pages, the intent is commercial or transactional. If it’s full of blog posts, videos, and AI-generated summaries, the intent is informational.
3. Map supporting content across the buyer journey
Very few purchases happen the first time someone encounters your brand, especially in B2B, where sales cycles routinely stretch to many months. That means your conversion-focused page needs supporting content to nurture people earlier in their decision process: comparison articles, how-to guides, and evaluative content that helps someone move from “I have a problem” to “I’m ready to choose a solution.”
Good supporting content doesn’t sell aggressively. It answers real questions honestly, which is exactly what earns trust (and, increasingly, citations inside AI-generated answers).
4. Document and track your keyword map
Keep a running record of every keyword you’re targeting, which piece of content owns it, the search intent, and the status of that content (planned, in progress, published). A simple spreadsheet works fine. This becomes essential once your content library grows past a handful of pages, since it prevents duplicate effort and keeps the whole team aligned on what’s already covered.
Practical tips for choosing the right keywords
Use free tools before paying for anything. Google Keyword Planner, Google Trends, and lightweight tools like Ubersuggest are enough to get a solid first pass at keyword data without spending anything.
Don’t chase volume for its own sake. A hyper-specific, lower-volume term that closely matches your offering will often outperform a broad, high-volume term that only loosely relates to what you do.
Go after realistic targets first. Check who currently ranks for a term before committing resources to it. If the results are dominated by major, well-established brands, a newer site is unlikely to break in. Look instead for narrower variations of the topic where the competition is thinner.
Balance short and long-tail terms. Short, broad “head terms” carry more volume but are far more competitive and harder to rank for, especially for newer sites. Longer, more specific phrases usually convert better and are easier to rank for, precisely because they signal clearer intent.
Factor in what’s already showing on the results page. Some searches trigger image carousels, video results, or featured snippets. If a term consistently returns video content, for example, a purely text-based article is fighting an uphill battle for that spot; consider producing video instead, or alongside it.
Study your competitors’ keyword targets, selectively. Understanding which terms competitors are targeting is useful context, but it shouldn’t dictate your own priorities wholesale. Just because a term matters to a competitor doesn’t mean it matters to your specific audience, and gaps they’ve ignored can be real opportunities.
How AI has changed keyword research
From exact-match phrases to semantic intent
Search engines no longer rely purely on matching the literal words in a query. Modern systems interpret meaning and context, grouping closely related phrasings together as expressions of the same underlying intent. In practice, this means you don’t need a separate page for every minor variation of a phrase (“best CRM,” “top CRM software,” “CRM tools”). A single, comprehensive resource that fully covers the topic tends to perform better than several thin pages each targeting a narrow variant.
From keyword lists to topic clusters
Rather than chasing a long list of loosely related terms, it’s more effective to build a cluster of content around a single theme: one central “pillar” page covering the topic broadly, supported by several more specific articles that link back to it. This structure signals topical depth to search engines and mirrors how AI systems tend to evaluate authority: by the strength of the overall topic coverage, not any single keyword match.
AI tools can accelerate research, but they don’t replace it
AI-assisted research tools can surface keyword ideas, related questions, and content gaps faster than manual research alone. That’s genuinely useful. But AI-generated keyword suggestions still need to be checked against real search data: does anyone actually search this term, is it realistic to rank for, and does it connect meaningfully to your business goals. Treat AI as a research accelerator, not a replacement for validating data.
Optimizing for zero-click and conversational search
As more queries get answered directly inside AI Overviews and chat-based assistants, ranking position alone matters less than it used to. To stay visible in this environment:
- Structure content so a single, clearly written section can directly answer a specific question
- Use question-and-answer formatting where it fits naturally
- Build genuine topical authority rather than shallow, keyword-stuffed pages
- Keep information accurate and current, since freshness and reliability appear to influence what AI systems choose to cite
Choosing keyword research tools
You don’t need an expensive toolset to get started. A workable stack for most beginners:
- Google Keyword Planner — free, and pulls data directly from Google; the tradeoff is that it requires a Google Ads account and tends to group similar terms together rather than showing granular detail
- Google Trends — free, and useful for spotting seasonality and rising topics, though it shows relative interest rather than exact search volume
- Ubersuggest or similar lightweight tools — good for quick keyword ideas and content angles, usually with a daily limit on free searches
Once a content program is generating meaningful traffic and revenue, it’s often worth investing in a more comprehensive paid platform (Semrush and Ahrefs are two of the most established) for deeper competitor analysis, backlink data, and more granular difficulty scoring.
How often should you revisit keyword research?
Treat it as ongoing, not a one-time project:
- Continuously, in a light way: stay alert to new topics, shifting language, and emerging questions from your audience
- Quarterly, more formally: re-check search volume, competition, and relevance for your existing keyword list
- Annually, as a full audit: reassess your overall keyword strategy against business goals, retire underperforming targets, and identify new topic areas worth pursuing
It’s also worth doing fresh research any time you launch a new product, enter a new market, or notice your audience’s needs have shifted.
Getting started
Keyword research isn’t a box to check once and forget. It’s the foundation that everything else in a content strategy sits on top of: what you write, who it’s for, and whether it has any realistic chance of reaching the people who need it. Start narrow, focus on intent over raw volume, and build outward from there.
