There’s been an increasing amount of chitter chatter in the SEO community about long-tail searches getting longer. Most of this seems to be tied to the rise of AI.
Let’s back up – the early days of search
If you rewind to the early days of search, people understood that computers struggled with natural language. Queries were stripped back to the essentials, short, direct, and deliberately free of anything that might confuse a search engine.
As search evolved, so did user behaviour. People became more precise, and long-tail queries emerged. Instead of broad terms, users started searching with clear intent, specifying things like colour, size, location, or service type.
AI influence on keyword length
Now, with AI in the mix, queries are stretching even further. The reason is simple: AI invites conversation. When a tool responds in a friendly, human way, people naturally mirror that tone. Queries become fuller, more conversational, sometimes even polite.
As more people interact with AI in this conversational way, it’s influencing how they use traditional search engines too. What we once called long-tail queries are starting to look more like the middle ground.
You could think of it like this: short, long, and now, ultra-long queries. Some estimates suggest these ultra-long searches make up around 70 per cent of all queries.
There’s also a device angle to consider. Typing on a desktop or laptop tends to encourage longer, more considered queries, while mobile searches may skew a little shorter. And voice search adds another dynamic into the mix. It’s estimated that the majority of voice searches are long to ultra-long.
So what do short and long-tail searches look like now?
- Short queries are still one to two words.
- Mid-length queries might be two to four.
- Traditional long-tail queries sit above that.
- But AI is pushing a new category, ultra-long queries, often 10 words or more.
Some of that extra length comes from so-called function words, words like “a”, “in”, “to”, and “for”. These act as the glue of natural language. In shorter searches they were stripped out, but in conversational queries they come back in, adding to the overall length.
What about search volumes and competition?
From a competitiveness perspective, the old rules still broadly apply. Short queries tend to have high volume and high competition. At the other end, ultra-long queries have lower volume and are less competitive.
In theory, that makes them easier to rank for. Hurrah! More importantly, they often signal strong intent. Someone typing a detailed, specific query usually knows what they want, which means the resulting traffic can be highly valuable. Hurrah again.
Use natural language to rank for long-tail searches in AI
Using longer, more natural phrases in your content helps it match how people actually speak and type, especially in AI searches. This makes it more likely that AI tools will also use your content in their answers. In addition, AI queries often ‘fan out’ to provide a really comprehensive answer by looking for related content, so this may also help content appear within other AI results too.
How to develop an ultra-long keyword strategy
Of course, there’s a practical limit. Individually, ultra-long queries have low volume, and trying to target every variation would be endless work.
The smart approach is in creating sensible groupings. Many of these queries share a common theme or intent, and can be addressed together within a single piece of content. When grouped effectively, their combined search volume can rival, or even exceed, that of a single high-volume short-tail keyword.
So how should SEOs respond?
The shift towards longer, more conversational queries doesn’t mean abandoning traditional keyword strategy. It means evolving it. Think less about isolated keywords and more about topics, intent, and how real people phrase their questions.
To find these queries, keyword tools are still useful. If you don’t have access to them, there are plenty of alternatives. People Also Ask results, Google’s autocomplete suggestions, forums like Reddit, Quora or industry-specific discussion sites, and even simple conversations with your target audience can all reveal how people actually search.
Google Search Console is another valuable source. It shows the queries your site already appears for. If you’re ranking for a mid-tail term, expanding that page to address related queries can be an efficient way to capture additional long-tail and ultra-long-tail traffic.
However you approach it, the underlying principle is the same. Write for humans, not algorithms. The closer your content reflects how people think, speak, and search, the more likely it is to perform, both in traditional search and in AI-driven results.