How Artificial Intelligence is Powering Google Search

Emma Grant

Head of SEO

Results-driven Emma heads up our SEO team. A champion of best-practice SEO strategies for search engine success that lasts and delivers best value, Emma’s work focuses on boosting clients’ online visibility with the ultimate goal of increasing sales. Emma applies her extensive skill and experience to create strategies that pay off quickly for faster results. She knows precisely what to do to reduce clients’ paid ad spends for greater profits, and how to boost organic leads for better long term return on investment. After close of business, when she’s not organising exciting travel adventures, you may hear Emma strumming classical guitar tunes or working on her jazz riffs.

Artificial intelligence in search

When you’re investing in an SEO campaign, it’s incredibly useful to know how Google understands the human intent behind search queries, and how it delivers results. Recently, the search engine has revealed its driving technologies, and they’re all centre on artificial intelligence and machine learning.

Google has recently published an outline of its major artificial intelligence (AI) systems in the context of Search. It’s a great insight into how its models work, and how they are becoming more and more sophisticated in terms of understanding the human intent behind search queries.

Let’s explore Google’s search technologies.


RankBrain was launched in 2015 and was Google’s first deep learning system. Still used today, it helps the search engine understand how words relate to real world concepts.

RankBrain uses machine learning to understand the user intent of long tail keywords so that it can deliver the most relevant results.

In the name of achieving top search results, we need to refine and improve our page content to ensure it matches user intent, making sure we directly answer the questions that are being asked.

Neural matching

It was 2018 when Google brought neural matching to organic search, and 2019 when it started playing a role in local search.

This technology looks at queries or pages as a whole rather than just keywords. It helps in understanding concepts that aren’t necessarily clear, which means Google can cast its net wider when looking for content that matches the search query.

Google says neural matching is a critical part of how it retrieves relevant content from a huge and continuously evolving information stream.


BERT stands for ‘Bidirectional Encoder Representations from Transformers’. Launched in 2018, it is very powerful, because it has the ability to understand search intent, cross-reference the context, and analyse relationships between words.

You’ll have experienced BERT when you’ve been researching something online and then later go to type in another query, and Google instantly comes up with search suggestions that relate to what you were previously looking at. It’s all about associating your previous searches with your current search. BERT really is artificial intelligence personified, and it’s great for user experience too.

Google says BERT plays a vital role in almost every English search query, doing a great job of ranking and retrieving, which are two of the most crucial tasks when it comes to returning relevant search results.


Launched in May 2021, Google Multitask Unified Model (MUM) is a pioneering AI-powered technology. It is designed to answer complex questions that don’t necessarily have straight answers.

Google MUM is a thousand times more powerful than BERT, and was designed to address the issue of having to enter multiple queries and conduct lots of searches when we are collating research information.

MUM is built on a Transformer architecture, just like BERT. But it is engineered not just to understand language, but also generate it. It is trained across 75 different languages and, because it is able to multitask, it has the power to develop a more comprehensive understanding of world knowledge and rich information than BERT and other previous models.

MUM is also multimodal. This means it understands information not just across text, but also images, and in the future it will work with video and audio too.

This technology is currently in the early stages of development, so is not yet being used for ranking or improving search results in the same ways as RankBrain, neural matching and BERT. But Google has said that MUM will be offered in the months ahead as a more intuitive method of searching, using a combination of text and images in the virtual reality-driven Google Lens.

For SEO expertise that’s up with the latest search technologies, talk to Figment

Here at Figment, we know the importance of staying on top of the latest search and ranking trends and technologies. It’s vital to evolve with Google and prepare for future shifts, so we can ensure our clients achieve and maintain top ranking positions for the online visibility they need to drive sales and ace their business growth goals.

Looking for a trusted, top rated SEO agency in London or Surrey?  To learn how we can help you grow your business with proven SEO expertise, please get in touch.

SEO results made simple

Request a Quote