Insights

SaaS Keyword Research in 2026: Why Strategy Now Starts With the Buyer, Not the Term

May 26, 2026

Picture the keyword plan sitting in your content folder right now. It’s probably a clean spreadsheet, sorted by funnel stage, with search volume and difficulty scores in neat columns. You did the work. You mapped intent. You built a content calendar from it, and you’ve been publishing against it for months. Traffic is up. And yet the demo requests haven’t followed.

If that sounds familiar, here’s the thing you need to hear. Your keyword list isn’t wrong. It’s just being asked to do a job it can no longer do on its own. For years, saas keyword research worked as a stand-in for strategy. You found the terms your buyers searched, you mapped them, and the plan more or less wrote itself. That’s stopped working on its own, and it’s worth understanding why before you build another quarter of content on top of it.

This guide won’t walk you through a step-by-step process. Plenty of those already exist, and you almost certainly don’t need another one. Instead, it’s about where to point your effort. The core argument is simple: in 2026, a strong saas keyword strategy starts with researching the buyer, not the term. Keyword research is still useful, but it has become one input into a bigger exercise rather than the exercise itself. Let’s look at what changed, and what you should do about it.

TL;DR

SaaS keyword research in 2026 still matters, but it’s no longer the strategy. The strategy now starts with the buyer. Here’s what that means in practice:

  • Buyers stopped searching in keywords. 84% of CMOs now use AI tools like ChatGPT, Claude, and Perplexity for vendor discovery, and 68% start there before opening Google. Queries have become full sentences, and AI agents research vendors on the buyer’s behalf.
  • The unit of optimization has shifted from the keyword to the buyer. What wins is the specificity of the match between a clearly defined buyer and your product’s positioning. Describe your product for “tradesmen running teams of 2 to 10 staff,” not just “field service management software.”
  • Answer three questions before you open a keyword tool: Who exactly is this buyer? What job are they hiring software to do? What makes your product the obvious fit for that exact person? Keywords come after, as expression, not discovery.
  • Cover all three funnel stages, but treat each as a buyer state. Top of funnel: be specific enough to be remembered. Middle of funnel: answer the sub-questions an AI agent asks during query fan-out. Bottom of funnel: recognize the buyers already describing themselves in search.
  • Build for scrutiny, not just discovery. When an AI Overview appears, buyers slow down, re-read, and compare listings before clicking. Specific, buyer-aligned content holds up under that scrutiny. Generic content doesn’t.
  • Keyword research still earns its place by surfacing real buyer language, revealing existing demand in Google Search Console, and structuring content into clusters. Measure pipeline and SQLs, not rankings.

Buyers stopped searching keywords

Before we get into recommendations, it’s worth being precise about what’s shifted. Two things changed at the same time, on two different surfaces. The first is where buyers do their research. The second is how they behave even when they do land on a traditional results page. Both point in the same direction.

Research has increasingly moved into AI assistants

Your buyers are increasingly starting their research inside a chat window, not a search bar. According to a 2026 Wynter survey, 84% of CMOs now use tools like ChatGPT, Claude, and Perplexity for vendor discovery, up from just 24% a year earlier. More striking, 68% of them open an AI tool before they open Google.

The pattern holds further down the buying committee, too. A March 2026 G2 survey found that 71% of B2B software buyers rely on AI chatbots for software research, and that 85% think more highly of a vendor when a chatbot mentions it in a recommendation. So this isn’t a fringe behaviour you can afford to watch from a distance. It’s how the majority of your market now works.

On top of that, buyers aren’t just chatting. They’re delegating. Autonomous agents now research vendors on a buyer’s behalf. As Discovered Labs describes it, an agent told to “find the top three incident management tools for a 200-person SaaS company” will query multiple sources, weigh them, and hand back a synthesised shortlist. The buyer never sees a results page at all.

Google itself has confirmed the trajectory. At its I/O 2026 event, the company explained that it rebuilt the search box because, in its words, “your curiosity doesn’t always fit into keywords.” It also started rolling out search agents that work in the background around the clock. When the largest search engine on earth tells you queries no longer fit into keywords, that’s a signal worth taking seriously. It’s the same shift we’ve written about in our breakdown of what visibility inside Google AI Mode now demands, where six and seven word queries have become a startpoint. Once “long-tail” phrases, these now represent a searcher starting to build context.

Even the Google results page became a slower, more deliberate place

You might assume all of this only matters off the traditional results page. It doesn’t. A May 2026 study published by Search Engine Journal, which analysed 846,000 Google search sessions, found that the results page itself has changed character. When an AI Overview appears, searchers stay longer, read more carefully, and scroll back up the page far more often to compare options before they click.

The study found something else worth noting. Without an AI Overview, the five search-intent types behave like distinct audiences. With one, they converge into a single, more deliberate audience. The presence of an AI Overview predicts behaviour more strongly than the type of search someone ran. In other words, the query is no longer a reliable label for a person. The query is now a description of a person and their situation, and keyword research that stops at the term is researching the wrong object.

From keyword targeting to buyer targeting

So if the keyword isn’t the unit of strategy anymore, what is? It’s the buyer. More precisely, it’s how sharply you’ve defined the buyer and how closely your product’s positioning matches them. Keyword research doesn’t go away in this model. It just gets demoted from “the strategy” to “the surface layer of a deeper exercise.” Let’s make that concrete, because it’s easy to nod along with and hard to actually do.

Specific positioning beats broad category coverage

Think about how most SaaS companies describe themselves. “Field service management software.” “Property management software.” These are categories, and a category is a crowded place to compete. Now compare those with sharper versions: “field service management software built for tradesmen running small teams of two to ten staff,” or “all-in-one property management software for UK landlords who want to stay MTD compliant.”

The second version of each describes a person, not a product type. And that’s the point. When you target a specific buyer, your keyword work doesn’t disappear. It gets better, because the terms you choose now express a real person with a real situation instead of a generic category label. This is also why positioning and search can’t sit in separate rooms.

Three questions that should come before your keyword tool

Here’s a practical way to apply this. Before you open a keyword tool, answer three questions. They look simple. They aren’t.

  1. Who exactly is this buyer? Not “marketing managers.” Which segment, which company size, which role, and which specific situation they’re in right now.
  2. What job are they hiring software to do? Describe the outcome they want in their words. A buyer doesn’t search for “expense software.” They search for how to automate invoice approvals.
  3. What makes your product the obvious fit for that exact person? This is the positioning claim your content has to earn, not just assert.

Once you’ve answered those, keyword research becomes the step where you express what you found. It’s worth saying clearly, though, that none of this means keywords are dead. AI systems still favour specificity and genuine expertise. What works now is depth, not the same generic blog post a buyer has already read on five competitor sites. Keyword research is demoted, not deleted. It just moves to a later spot in the sequence.

Where to focus: keyword strategy across all three funnel stages

With the buyer at the centre, the funnel still matters, but you should think about it differently. Each stage isn’t a keyword bucket. It’s a buyer in a particular state of certainty. And because AI now sits between your buyer and your content at every stage, the job each piece of content has to do has shifted. Let’s walk through all three.

Top of funnel: be specific enough to be remembered

Top-of-funnel content targets a problem-aware buyer who hasn’t started comparing tools yet. Here’s the honest take. AI now answers many of these informational questions directly, so a generic “what is X” article gets absorbed into an AI Overview and earns you nothing. That traffic is largely gone.

But top of funnel isn’t worthless. It’s changed. Agents increasingly anticipate that a user will need a tool, so an informational answer often arrives with a product recommendation attached. The Search Engine Journal study adds a useful detail here: when even casual searchers are pulled into slow, back-and-forth comparison, vague content doesn’t survive the scrutiny. So your top-of-funnel job is to be specific enough about the problem that an AI model connects your brand with the solution to that exact problem, not the broad category around it.

Middle of funnel: answer the questions an agent will ask next

Middle-of-funnel content serves a solution-aware buyer who knows they need a tool and is now weighing options. This is “best X for Y,” category, and comparison territory, and it’s where the buyer-first reframe matters most.

Tie your research directly to how agents behave. During what’s known as query fan-out, an agent breaks a buyer’s question into follow-up sub-questions: pricing, integration depth, migration effort, support response times. Then it pulls from whichever source answers each one most cleanly. So your middle-of-funnel keyword research isn’t a list of comparison phrases. It’s an inventory of the questions your buyer and their agent will ask, paired with a commitment to answer every one of them concretely, right there on the page.

Bottom of funnel: recognise the buyers describing themselves

Bottom-of-funnel content meets a buyer at the decision stage. These are your competitor, alternative, and comparison queries, and they deserve to be the core of your strategy rather than an afterthought. They’re the highest-converting, lowest-competition, fastest-to-rank terms in B2B SaaS.

Look closely at a query like “best CRM for personal injury law firms under 20 attorneys.” That isn’t really a keyword. It’s a buyer definition, written by the buyer. Your job at this stage is to spot which specific buyers are already describing themselves in search, then make sure your positioning meets them exactly where they are. The mistake to avoid, across all three stages, is bunching your keywords at one level. The goal isn’t three separate buckets. It’s content that recognises the same buyer at three different moments of certainty.

Why your search presence must be built for scrutiny

There’s one more reason specificity wins, and it shows up in how people behave on the results page itself. The same Search Engine Journal study of 846,000 sessions gives us a clear picture, and it reinforces everything above from a fresh angle.

Buyers now read, revisit, and compare before they commit

When an AI Overview is present, buyers don’t scan and click. They read, they pause, and they scroll back up the page to re-compare listings they’ve already passed. That back-and-forth scrolling is a signal of active comparison and reconsideration. It means your listing gets looked at more than once, and that it gets weighed against competitors in the same glance.

This raises the stakes for your search result preview. Your title tag and meta description now carry more weight, because buyers pause over the snippet and return to it rather than reflexively clicking the first result. The study’s own conclusion puts it well: the brands best positioned in this environment are the ones whose search presence is built for scrutiny, not just for discovery.

Connect that back to positioning. A vague, generic listing might pass a quick scan, but it falls apart under a slow, deliberate re-read next to a sharper competitor. This is the same lesson as the positioning point earlier, except now you can see it in behavioural data rather than just argue it in principle. Specific, buyer-aligned content isn’t only easier for an AI to recommend. It also holds up when a real person stops to scrutinise it.

Where keyword research still earns its place

By now it might sound like keyword research belongs in a museum. It doesn’t. Demoting something isn’t the same as discarding it, and keyword research still does real work once you’ve put the buyer first. Here’s where it genuinely earns its place.

It surfaces the language your buyers actually use

Keyword research is still one of the best ways to capture how your buyers really talk, as long as you validate it against voice-of-customer sources. Sales calls, support tickets, G2 and Capterra reviews, and community threads all show you phrasing that a tool will never surface on its own. That matters, because the reference articles in this space agree on an uncomfortable point: the highest-value, most specific B2B terms often show “zero volume” in tools like Ahrefs or Semrush. Treat a tool as one input, not the final verdict.

It reveals demand you can already see

You don’t have to guess your way to opportunity. Google Search Console shows you queries where you already earn impressions but few clicks, and each one is direct evidence of a buyer-language gap you can close. 

The natural starting point is auditing what your site already ranks for. From there, keyword research also gives you structure. Clustering related buyer questions keeps your content connected and helps you avoid the cannibalisation that comes from chasing tiny term variants one content piece at a time.

One last point on measurement, because it ties the whole thing together. Rankings and traffic are diagnostics. They tell you the machinery is working. They are not the goal. The goal is pipeline, qualified leads, and organic-influenced revenue. Keep your eye on MRR rather than rankings, and judge keyword research by whether it points content at buyers who actually convert.

Research the buyer. Let keywords express what you find.

If you take one thing from this guide, make it the sequence. In 2026, saas keyword research isn’t something you stop doing. It’s something you stop doing first. Define the buyer and the positioning, then let keyword research express that definition across all three funnel stages.

That’s the real shift behind a modern saas keyword strategy. As search becomes agent-mediated, the brands that win aren’t the ones holding the longest keyword list. They’re the ones an AI can confidently recommend to a narrowly defined buyer, because their positioning is specific and stays consistent everywhere a model looks. The keyword list still has a job. It just answers to the buyer now.

So before your next planning cycle, resist the urge to open a tool first. Start with the person you’re trying to reach, get specific about who they are and what they’re trying to do, and treat organic search as a revenue system that supports your go-to-market motion. Do that, and your keywords will finally point at the people you actually want to talk to.

Related Posts