A couple of weeks ago, I outlined my case for AIO (artificial intelligence optimisation) in an article titled From SEO to AIO.
When I say AIO, I’m borrowing a term that refers to optimising LLMs, however, I believe the term will gain a double meaning.
In my use of the term, AIO essentially means the same thing as SEO (search engine optimisation), but for search tools that rely on AI technology.
My basis for this is that AI technology will change how search engines work and how they serve information to audiences.
So, to outline how I think this will work, it’s time to take a look at how classic search works versus how I believe an AI-driven search engine will operate.
Comparing search models
Classic search, ie. how Google works
At a very (very very) basic level, Google works like this:
- Googlebot (Google’s web crawler) travels the web finding websites
- When Googlebot visits a website, it finds web pages
- On each webpage, Googlebot discovers content
- Googlebot figures out what the content is communicating
- It returns to a big virtual filing cabinet (the index) and files web pages and content away in the correct folder in the cabinet
The details, in reality, will be much more nuanced, however, the general approach outlined above describes how Google goes about discovering content, understanding it, and storing it in an index for future retrieval.
The key takeaway is that Google crawls the web to find web pages and content to add to its index.
When a user searches, Google returns to its index, finds matching web pages or content, and displays the results to the user.
How AI-driven search will work
AI-driven search will be able to handle information separately from its delivery mechanism.
We’re yet to truly see a ubiquitous search tool that utilises an AI-driven model in the way I’ll be referring to. Google certainly uses machine learning (ML) to figure things out, and the likes of Alexa and Siri use a form of AI to work. What I’m referring to, though, is a generative engine that replaces classic search.
I believe that AI-driven search will view information in an entirely different sort of way.
Whereas a classic search engine finds information in the form of content, which in turn is found in the form of a webpage, AI-driven search will be able to handle information separately from its delivery mechanism.
Think of it like this; if you learn a new fact when reading a book, you have gained that knowledge as a discrete piece of information. The information now resides in your memory.
That same piece of information lives within the book, but it also lives within your memory. At the same time, it also lives in the mind of the author and other media; it could be in a video, a webinar, an online document etc.
I believe AI-driven search will treat information in a similar way.
Every piece of information the search tool learns will become a part of its knowledge base.
Now, when a user searches, the engine doesn’t need to deliver a link to a webpage, it can deliver a piece of information in whatever format suits the situation best.
Why AI-driven search will use information differently
Everything can be reduced to entities and relationships.
The quick brown fox jumps over the lazy dog.
The quick brown fox is an animal and, therefore an entity.
So too is the lazy dog.
The fox’s relationship with the dog is that it jumps over it.
So what we have is that one entity has a relationship with another entity.
Pretty much everything can be expressed in this way.
This will be critical to how the future of search works. A search engine will rely on how it understands entities and relationships. This understanding will determine what it communicates with its users.
How this will turn SEO into AIO
In short; marketers won’t need to optimise websites and web pages anymore, they will need to optimise knowledge and information instead.
Where once you had a webpage to describe your core service, you’ll now rely on ‘sources of truth’ to ‘train’ search engines on your core offering.
Your website will still play a key role; it’ll be your main ‘source of truth’. It will be joined, however, by other sources of truth such as Wikipedia, social profiles, and trusted third parties.
A search engine will switch from crawling websites to find pages to crawling the web to find information it can add to a knowledge graph.
It will work like a big brain, soaking up knowledge and using its best endeavours to figure out what knowledge it can trust and what information comes from a skewed source.
It will link what it has learned about an entity in one place with complementary information it has found in another place.
Marketers won’t need to optimise websites and web pages anymore, they will need to optimise knowledge and information instead.
What do content owners need to do?
I believe that we’re moving towards a time where content will need to become ‘responsive’.
In the same way that a website responds to the device size it’s viewed on, content will need to become responsive to context.
For example, an article about how to find your old pension pots can be repurposed into:
- A bullet point list of key steps
- A two-sentence abstract
- A short animation outlining the process
- A full-page guide
- A podcast episode
Depending on when the information is required, an AI-driven search tool will use its knowledge to serve the correct format.
If you are in the car talking to it, it may read out the abstract. If on your phone, it might add the bullet point list to your to-do app. If you are on your desktop, it may play the video for you.
In each case, it will back up claims by providing a source for the information (ie. you and your brand).
Responsive information will be the key to a successful AIO strategy.
