Insights

I remember when search engine optimisation was considerably simpler.

If you ran a coffee shop in Mandurah and wanted people to find you online, one of the search terms you cared about might have been:

coffee Mandurah

Two words. What you offered and where you offered it.

Nobody walked up to another person and said, “coffee Mandurah”, but that's how we learned to communicate with search engines. We stripped what we actually wanted down to the handful of words we thought they would understand.

Then search got smarter.

Google became better at understanding context, mobile devices knew where we were, and location stopped needing to be explicitly included. Instead, we searched:

coffee shop near me

The search became shorter while the technology behind it became considerably smarter.

Now we're watching that behaviour change again.

From keywords to actual questions

Voice assistants such as Siri and Alexa got us comfortable with speaking to technology rather than carefully constructing a search query. Generative AI has taken that much further.

If I'm travelling and looking for somewhere to work for an hour, I don't need to search “coffee shop near me”. I can explain what I actually want:

“Where can I get a decent coffee within a ten-minute walk that's open now, has Wi-Fi and somewhere quiet enough to work for an hour?”

Twenty years after we learned to simplify our questions for search engines, we're putting all that detail back in.

In some ways, long-tail search is making a huge comeback. Except this time we're not constructing longer keyword phrases for a machine.

We're having conversations.

Finding a coffee shop is easy. Choosing the right one isn't.

That longer question presents a very different challenge.

The system needs to know where I am and how far I'm prepared to walk. It needs current opening hours, whether somewhere has Wi-Fi and potentially information about its facilities, accessibility and popularity.

It then has to bring that information together and decide which places actually fit what I've asked.

That's why a recent announcement from Mapbox caught my attention.

We've used Mapbox in custom development projects for years, particularly where location and geographic data are important.

Its new Places API provides detailed information about more than 250 million places worldwide.

The number isn't the interesting part.

It's the context being attached to those places.

Alongside a name, address and coordinates, a place can include opening hours, contact information, photos, accessibility information and other attributes, as well as its relationship with surrounding buildings and neighbourhoods.

A coffee shop is no longer simply a pin on a map. It becomes a collection of information that software can use to understand what that place actually is.

And that's where AI becomes interesting.

Being found isn't the same as being understood

When somebody searched “coffee Mandurah”, a search engine needed to find pages that matched those concepts.

With “coffee shop near me”, it could add location and intent.

Now, when somebody asks a detailed question, they're asking the technology to make a judgement about which business best fits their requirements.

That judgement requires context.

This is why looking at AI search as simply another place where your business needs to rank misses the bigger change.

Being indexed matters. Being understood matters more.

Your digital identity is bigger than your website

Your website remains one of the strongest places you control for explaining what you do, who you help and where you operate.

But it doesn't exist in isolation.

Information about an organisation can sit across its website, Google Business Profile, mapping platforms, directories, reviews, social profiles, structured data and other sources.

Together, those sources create what we've increasingly been describing as a business's digital identity.

As questions become more specific, the systems answering them need more context to decide whether a business is relevant.

That makes the consistency and quality of that wider digital identity increasingly important.

SEO isn't disappearing. Search is evolving.

SEO has supposedly died countless times already. AI isn't suddenly making websites, Google or traditional search irrelevant either.

What's changing is the breadth of what search optimisation needs to consider.

With conversational search, there could be hundreds of ways somebody describes a situation for which your business is the right answer.

Trying to optimise individual pages for every possible phrase doesn't make sense.

Building a clear, authoritative and well-structured digital representation of the business does.

There's an interesting reversal happening here too.

Long-tail searches traditionally had lower search volumes but often revealed much more about what somebody actually wanted. AI and voice search remove much of the reason to shorten those questions.

The wording becomes less predictable while the intent becomes more precise.

That's a significant change.

It changes what we can build too

For developers, this isn't only about search marketing.

Richer data combined with AI opens up entirely new digital experiences.

A tourism platform could answer questions based on where a visitor is, what they're interested in and what's actually available nearby.

An accommodation website could help a guest find somewhere within walking distance for dinner that's still open when they arrive.

A property platform could provide meaningful context around a location rather than simply displaying pins on a map.

The individual technologies aren't necessarily new. What's changing is how effectively we can connect them and how naturally people can interact with the result.

Would AI understand your business?

Back when “coffee Mandurah” was the target, the challenge was relatively easy to visualise:

Does Google understand that this page is about coffee in Mandurah?

Today there's a better question.

If someone describes a problem or requirement for which your business would genuinely be a great solution, is there enough information available for an AI system to work that out?

Would it understand exactly what you offer, where you operate and who you're particularly suited to helping?

Would the information it finds across different sources reinforce that understanding or contradict it?

Ultimately:

If AI had to explain your business to a potential customer, would it get it right?

For more than two decades, we've adapted what we publish online to help search engines understand us.

Now those systems can understand considerably more.

The opportunity is making sure we're giving them the right information to work with.