How we build · Included as standard
Every site we build
is built to be found.
Search is splitting in two. Some customers still scroll a page of blue links. A growing number ask ChatGPT, Claude, Perplexity or Google’s AI overview, then act on the answer without visiting a single site. Most websites are built for the first and are invisible to the second.
We do not sell this as a service, because it should not be one. Being readable by a search engine and by an AI assistant is part of how a site gets built. So is working properly on a phone. It is on every project at no extra cost. This page is simply the standard we hold ourselves to, published so you can hold us to it.
Part of Nexus Cloud
This page describes how we build. The other half is what happens afterwards. Everything we build runs on Nexus Cloud as a managed service, so the standards described here do not decay once the invoice is paid.
The shift
If the answer never links to you, the ranking did not matter.
An AI assistant answering “who builds property websites in Abu Dhabi” does not read your homepage the way a visitor does. It reads structured data, machine-readable summaries and question-shaped content, and it weighs how often independent sources say the same thing about you.
Sites that give it nothing to parse get left out of the answer entirely, no matter how good they look.
What every system includes
- Schema.org structured data for your organisation, services and location
- An llms.txt summary at the site root, an emerging convention that costs nothing to publish
- Crawler access decided per bot, separating search indexing from model training
- Question-shaped content with FAQ markup
- Clear entity signals: who you are, where you operate, what you do
- Classic technical SEO, sitemaps, titles and metadata
How it works
Four things that make a site citable.
Be readable
Structured data states your facts in a format machines parse without guessing: services, location, contact, service area.
Be summarisable
An llms.txt file gives AI crawlers a clean summary and a map of your pages, instead of leaving them to infer it from markup. It is a young convention rather than a guarantee, which is why it is included rather than invoiced.
Be reachable
Many sites block AI crawlers by default, often without knowing. We check and open access deliberately.
Be quotable
Content written as real questions with direct answers is what assistants lift into a response.
Straight answer
What this does and does not do.
Nobody can make an AI assistant recommend a business. There is no setting for it, and any agency promising one is selling you something that does not exist.
What is controllable is whether these systems can read you, understand you and cite you. Sites that are parseable and consistent get included in answers. Sites that are not get skipped.
We also tell you the part that is not on your website: the mentions, reviews and directory listings that decide whether an assistant treats you as corroborated. We map those out with you rather than pretending the site alone is enough.
The difference
A traditional website.
And one built to be found.
Most web design work in this region still stops at the point the site looks good. These are the specific, checkable differences in what we hand over.
| Capability | Typical website | Built by ProcureNexus |
|---|---|---|
| Structured data | None, or a plugin default that describes nothing specific | Organisation, service, location and inventory described explicitly |
| AI crawler access | Often blocked by default, usually without the owner knowing | Checked and opened deliberately, and documented |
| Machine-readable summary | Not present | llms.txt published and kept in step with the site |
| Content structure | Marketing copy written for a brochure | Real questions with direct answers, marked up so they can be quoted |
| Listings and products | Rendered for human eyes only | Published with structured data so each item can be read individually |
| After launch | Handed over and left to age | Monitored and updated as platform requirements change |
| Hosting and upkeep | Your responsibility to remember | Included in the managed service |
Every row above is something you can verify on this site before you hire us. Open our robots.txt, our llms.txt, or run any page through a structured data validator.
Common questions
How this works, in plain terms.
How do I make my website discoverable by AI search?
It is not one setting. It is a stack of things that have to be correct together, and then maintained. Structured data describing your business and your inventory. A machine-readable summary at the root of your site. Deliberate crawler permissions. Content written in the shape of questions people actually ask. Consistent signals about who you are and where you operate. Get one layer wrong and the rest stops counting. This is built into every site we run, as standard, so you do not have to learn any of it or keep up with it as the systems change.
Will my products or listings be visible to ChatGPT and Claude?
That is the point of building it this way. Someone asks an assistant for an apartment for rent in a specific area, or a supplier for a specific product. The assistant can only work with what it is able to read. If your listings sit in markup no crawler can interpret, you are not in the running. We publish your inventory with structured data that describes each item properly, which makes it eligible to be read, understood and cited. Nobody can promise a particular answer on a particular day, and you should be careful of anyone who does. But being readable is the difference between being a candidate and being invisible.
What does a business lose without this?
Every enquiry that goes to whoever was readable instead. The traffic that used to arrive from a search results page is increasingly being answered before the click, and the businesses named in those answers absorb the demand. It is the same shift that happened when companies without a website lost ground to companies with one, except it is moving faster. The cost does not show up as a bill. It shows up as enquiries you never knew existed.
Do I need to understand any of this?
No, and that is deliberate. You tell us what you sell and where you operate. We handle the structured data, the crawler access, the summaries, the content structure and the monitoring, and we keep it current as the platforms change their requirements. Most clients never see the technical layer at all. That is what they are paying for.
We already have a website. Can you fix it instead of rebuilding?
Usually, yes. We audit what you have, check whether AI crawlers are being blocked, test what structured data exists, and tell you plainly whether it is worth upgrading or replacing. Plenty of sites need a retrofit rather than a rebuild, and we will say so when that is the case.
The evidence
Not our opinion. Measured.
A peer-reviewed study from Princeton, Georgia Tech, the Allen Institute for AI and IIT Delhi tested nine ways of writing a page against generative engines across 10,000 queries. Four changes moved the needle.
| Change to the page | Measured lift in AI visibility |
|---|---|
| Direct quotations from credible sources | about 40% |
| Relevant statistics instead of adjectives | about 38% |
| Citing the sources behind a claim | about 30% |
| Clearer, more fluent writing | about 29% |
| Keyword stuffing | Little to no improvement |
Two things in that table are worth sitting with. The tactic most often sold as SEO, packing a page with keywords, did close to nothing. And the gains were largest for sites that were not already ranking near the top, which is exactly the position most businesses are in when they come to us.
Source: GEO: Generative Engine Optimization, Aggarwal et al., published at KDD 2024. We have linked it so you can check it rather than take our word for it.
The standard, not a sales pitch
Ask any studio to show you this page.
Everything above is checkable before you hire anyone. Open a site’s robots.txt. View a page with JavaScript switched off and see what is left. Run it through a structured data validator. If a studio cannot show its own site passing those tests, it is fair to ask why yours would do better.