Using AI to improve on-site search in WordPress

Many WordPress websites technically have search, but it rarely works well enough to genuinely help visitors.

On older websites, search is often little more than a keyword match against page titles or content. If the visitor doesn’t type the exact words used on the page, the result is either irrelevant or empty.

At the same time, visitor expectations have changed.

People are now used to asking questions in natural language, the same way they would ask Google or ChatGPT. Instead of typing a keyword, they ask things like:

• “How much are tickets for 2 adults and 2 children?”
• “Can I bring a dog?”
• “Is parking available on site?”

Traditional WordPress search struggles with these kinds of queries. It expects keywords, not questions.

This creates real problems for organisations with complex websites:

• Visitors can’t find key information quickly
• Contact forms become the “default search engine”
• Phone calls and email enquiries increase
• Support teams spend time answering questions already explained on the website

This is especially common on websites that have grown over many years. Content accumulates, navigation becomes harder to follow, and the user journey becomes less obvious.

The result is friction. Visitors know the information exists somewhere on the site, but they struggle to find it.

Instead of forcing people to navigate menus or guess the right keyword, AI allows visitors to simply ask a question and get the answer immediately.

AI-powered search changes how visitors use your website.

Instead of clicking through multiple pages or trying different search terms, they can ask a full question in their own words and get a relevant answer straight away.

That means less friction, less frustration, and a much faster route to the information they need.

For organisations with larger, older, or more content-heavy websites, this can make a huge difference. Visitors no longer need to understand your navigation, your page structure, or the exact wording used on the site. They just ask.

It also reduces pressure on your team. When visitors can find answers themselves, fewer routine questions come through email, phone, social media, or contact forms.

The result is a better user experience for visitors and fewer avoidable enquiries for your organisation.

After implementing AI search with Holp, support enquiry volumes dropped significantly as visitors were able to find answers themselves.

Comparing 2025 to 2024, Eureka! saw a significant drop in enquiries across all platforms:

Total enquiries

45,735 → 32,201
30% reduction

Breakdown:

Phone calls
25% reduction
(5,900 fewer calls)

Emails
30% reduction
(5,700 fewer emails)

Social media enquiries
55% reduction
(1,800 fewer messages)

Instead of contacting staff, visitors were able to find answers directly through the website. Freeing staff up to work on other important tasks.