Google Continues To Improve Search AI

To keep on top of SEO it’s important to keep a close eye on what Google is planning. Last month, Google’s senior vice-president of search, John Giannandrea, spoke to Leslie D’Monte from Live Mint about the increasing role or artificial intelligence in search. From the exclusive and in-depth interview we have gained some vital insights into what is on the SEO horizon.


In less than 20 years Google has gone from processing 10,000 searches per day to 40,000 per second (around 3.4 billion a day). Both Google and the Internet has changed immeasurably in the last 18 years; when Google started there were around 2 million websites, today there are over 860 million. What is maybe ever more shocking, is that there are now over 3 billion Internet users – almost half the world’s population is online. Most use Google every day.

This has all posed massive challenges for Google. It can no longer rely solely on a group of search engineers monitoring trends and adjusting algorithms manually to deliver the most relevant search engine results. The solution is a combination of artificial intelligence and machine learning.

Before joining Google six years ago, John Giannandrea was the CEO of Metaweb Technologies Inc., which Google acquired. He is now head of Computer Science Research and Machine Intelligence, which is driving change at the heart of Google.

Google has already made advances in key areas of AI, such as voice and image recognition, natural language processing and translation, but now it is looking to utilize AI to improve the way it ranks websites.

Giannandrea explained what this means for Google, and ultimately Google’s customers: “Machine learning improves the products we offer everyone, and makes possible what was impossible just a few years ago. You can talk to the Google app on your smartphone. Thanks to deep learning, you can improve speech recognition …. and natural language processing (NLP) helps understand what you mean. With neural networks, Gmail now blocks over 99.9% of spam, including some never-before-seen spam.”

Google’s ability to block email spam must also be connected with its ability to identify web-spam. We have seen how algorithms such as Panda and Penguin have massively reduced the amount of web (search) spam on the Internet, but something that Google so far has been very cagey about revealing is how much of this is governed by machine learning, and how much is down to manual programming.

Some of the biggest changes at Google in the last few years is the increasing importance of Knowledge Graph, which was why Google purchased Metaweb Technologies.

We can be sure that Google Search will continue to focus its attention one these two core areas: improving the quality and breadth of instant results, aka Knowledge Graph, and developing faster and more reliable machine learning algorithms that will eliminate all web spam and help rank the best businesses and website at the top if the SERPs.

Google has certainly made some staggering process with algorithms such as Humming Bird and RankBrain. Giannandrea, in a somewhat understated comment, said that Google has got “a good bit smarter, no longer relying on just matching keywords”. We have seen a lot of evidence of this in the last year alone – already many of the old SEO techniques that helped springboard companies to the top of search for the most important keywords are no longer working as well as they once did.

How RankBrain Has Changed Google

Giannandrea also provided some vital insights into RankBrain, and said that it has “helped Google understand long or complex questions better. And the Knowledge Graph ties it to real-world people, places, and things”.

It’s is still a rather grey area in the SEO world – we all know that it is there, and that it is huge, but what is Google Rankbrain? Nobody really knows what it does or how it works! This poses one of the biggest challenges that SEOs have seen for years, and there is no doubt that this is completely deliberate on Google’s part. Google does not really like us SEOs – it wants to have complete control of the search index, it does not want to have to rely on SEO professionals to help businesses rank well.

The most recent update, which is not currently linked directly to search, is Allo. But as Giannandrea is talking about it, we can be sure that the data gather from the Allo app (a chat app) is going to be used to further improve the artificial intelligence within Google. Giannandrea explained: “As we develop the Google Assistant, we are using all our experience building Search, and extending it to all kinds of other things you need to get done in your daily life. We’re still early on in this journey.”

So the future of search is more artificial intelligence, more machine learning, and more harvesting of user data to feed into these systems. SEO is becoming more challenging, but ultimately, so long as best practices are followed, a well-designed website with good quality and search optimised content, will perform very well.

One Reply to “Google Continues To Improve Search AI”

  1. Jay patel says:

    It’s fantastic post about search AI. Thank you.

Comments are closed.