How To Use AI To Get The Best Keywords For SEO

Top performing keywords make top performing websites. Machine Learning will find the best keywords for Google search.


It is not uncommon to hear the phrase ‘Just Google it!’ in the middle of a light-hearted debate with friends, or even ‘Ask Dr Google’ when chatting about health. Searching for information, and in many cases quite specific information, is a day-in and day-out process for the modern human being.

Searching typically requires keywords, but there are different kinds of keywords and combinations of keywords that any business operating in the online space needs to be aware of in order to convert searches into sales.

‘Google has already received 360 billion searches this year and counting’

Impact Plus 

Think of a person wanting to get into trail running. A ‘normal’ keyword or search term may be ‘trail running tips’ or ‘best trail running shoes’. In these instances, the newbie trail runner is wanting to get a general sense of trail running and the broad range of shoes suitable for off-road running.

The person is out to inform themselves: they will probably read a few articles; browse through different brands and models of shoes. There isn’t an intention to act behind the search – it’s like browsing in a bookstore, with no particular book in mind to buy.

It’s all well and good for a website’s SEO to have functioning keywords like these that get people clicking through to the site for information, and a quick SEO analysis will say your traffic is looking good. But if the site has 1 000 page visits but few meaningful interactions that result in sales, those keywords are not doing the business any favours. This is where performance keywords really shine.

Browsing Vs. Intentional Search

A performance keyword or search phrase is usually a string of words (long-tail keywords) that have a specific and often localised intention behind them.

Going back to trail running, after a bit of initial research the newbie trail runner knows more about running and what kind of shoes they should be looking for. This is when their search will get more specific with keywords like “Best beginner trails to run in Portland” or “Waterproof Gore-Tex trail shoes on sale”.

‘A carefully tuned SEO strategy based on effective keywords can mean less marketing spend with great results, maximising your ROI’


These very specific short- or long-tail keywords are performance keywords, as they indicate that the newbie trail runner is ready to hit the trails ASAP, and that they are now wanting to spend money on waterproof trail shoes for their runs around the Portland area. These visitors are primed for action and ready to spend, and your website needs to be able to attract them with relevant performance keywords in order to convert their active interest into sales.

The Performance Keyword Challenge

The difference between ‘normal’ and keyword performance is easy enough to understand. Getting performance keywords right is far trickier because of ever changing search trends.

While one could sit and think of a handful of performance long-tail keywords and have those embedded in the website, one must consider the diversity of the world and how different groups of people use distinctive terminology, slang words or colloquialisms in their Google searches.

Generationally, people will use different terms, too. Think of how a competitive female in her mid-20s would search for trail running shoes versus a male in his mid-60s looking to run for pleasure. Does the trail running shoe business want market share from both of those demographics? Yes! But in order to get business from both demographics, the website owner needs to be on top of the latest search trends for all age groups, genders and experience levels.

‘15%-20% of searches on Google have never been performed before’

Code In WP 

Keyword-based terminology, slang, colloquialisms, demographics… That’s an almost impossible list to stay on top of, especially if you’re only human. This is where ATRICA’s Artificial Intelligence (AI) comes into play: the AI stays on top of all keyword-based search trends relevant to your website’s business, including local and mobile-first search, and it does so faster and more efficiently than any human can.

It gets better: ATRICA’s AI also optimizes for semantic and entity search. Put simply, semantic search is a person’s intention behind the keywords they input into the search bar, which the Google search engine cleverly understands based on the user’s search history (i.e. what does the searcher typically like/not like).

Entity search is whether the website’s concept matches the user’s search. What we mean by concept is literally beyond words: the search engine takes into consideration images, topics, trends and other data making up a general concept that may be a good match to the user’s keywords (learn more about semantic and entity search here.

If semantic and entity search sounds complex, it’s because it is. So why spend the time trying to figure it out when you can get an AI to do it for you?

Using AI For SEO Keywords

ATRICA’s AI-driven software starts off by inputting 200 baseline keywords called Anchor Keywords (AKs) onto the website’s back- and front-end.

ATRICA’s AI is then ‘let loose’ to start monitoring the website’s best performing AKs based on Google search behaviour and competitor websites. The AI then hones in on the best performing AKs and, based on these, begins generating dozens upon dozens of Performance Keywords (PKs) for the website.

‘AI…allows us to make sense of an increasingly large amount of data in real-time’

AI Multiple

Without the website owner losing even a wink of sleep, ATRICA is monitoring performance keywords, suggesting better-performing keywords and implementing these on the website. Each day that ATRICA is on the job, it generates better and more relevant PKs that enable intentional action-orientated clients from all demographics to find the website.

And the non-performing PKs? ATRICA discards these into The Keyword Graveyard where they belong. It’s brutal, but if PKs are not doing their job they need to be tossed out, and quickly.

Making Local Search Trendy

ATRICA also optimizes top-performing PKs for local search, which is crucial for many businesses. And it does this automatically, even if you haven’t included localised search terms in your original 200 AKs.

‘Where to buy’ and ‘near me’ mobile queries have grown by over 200% in the past two years’

Impact Plus 

For example, the Portland-based trail running shoe shop may have inputted more generalised AKs, such as: trail running shop, zero-drop, sustainable materials, ultra-marathon, 50 Miler, 100 miler, trail running socks, shoe inserts, lacing systems, pronation…and so the list will continue.

If ‘trail running shop’ is a top-performing AK, ATRICA will automatically start generating and implementing PKs for localised search terms, such as “trail running shop near me” or “top 5 trail running shops in Portland”. And so the Portland-based trail-running shop will benefit from local searches that, statistically speaking, convert into more business than generalised searches.

Using AI To Hone In On Serious Customers

This is the beauty of AI-driven software; it inherently fine-tunes its own intelligence, learning from every ripple in the world of Google’s search algorithm.

ATRICA’s ability to zero in on performance keywords to attract clients who are ready to exchange money for services and products are second to none.

Which really means one thing: your ‘shoes’ will wind up on the feet of that trail runner, and if they’re happy with the service, you would have gained a return customer as well as their referrals.