What WordPress AI Really Builds: 9 Prompts, 9 Unedited Results, and What the Good Ones Have in Common

While many block plugins incorporate AI features, most tend to market them in a similar fashion: simply enter your prompt, and receive a complete page. They seldom disclose the number of prompts it required to achieve the displayed outcome.

For this article, I opted for a more open methodology by utilising the Section Generator in the free Otter Blocks plugin. I created nine sections for a WordPress site using a single prompt for each, rather than continuing to fine-tune them endlessly. I’ll be sharing screenshots of the outcomes here.

Some results were satisfactory enough for publication, while others were not up to par. The variations seemed to stem more from the content of the prompts than their phrasing. Let’s delve into the details.

Before You Begin: Establish Your AI Connection

Otter Blocks works harmoniously with the WordPress block editor, making it simple to access the Section Generator, just like any typical WordPress block. Additionally, it doesn’t depend on AI credits, unlike many other AI builders, which is another advantage.

Here’s the process: you describe what you need for a section, Otter builds it using its own blocks, and afterwards, you can modify everything as required. It’s as straightforward as that, but you must connect to an AI provider before you can begin generating sections. Here’s how:

• From the WordPress dashboard (i.e., wp-admin), head to Settings → Connectors.
• Link up a supported AI provider.
• Return to Otter Blocks → Settings → AI and save your provider settings.

And there you have it. Once connected, you can open the block editor, add the AI Block, and start generating your very first section.

For more detailed instructions on the initial setup, take a look at the AI documentation provided by Otter Blocks.

Now, let’s examine the sections I created and see how Otter responded to each prompt.

Where It Begins: Layout Achieved, Text Needs Improvement

The first section I constructed was a testimonials section, and it had the shortest prompt of all. Primarily, this example illustrates what the generator can and cannot achieve.

Prompt: A testimonials section featuring three customer quotes and their names.

We need to assess two components: the layout and the content itself.

The initial part performed admirably. I received three well-spaced cards, each featuring a quote above a corresponding name, and a good offset from the section above. In terms of layout design, it was a success.

Conversely, the text quality was a different matter:

“The entire process feels considered from beginning to end. We quickly found our footing and delivered superior results with significantly less hassle.”

It was standard, typical AI filler.

I specified a structure with three quotes and names, and the output consisted of a framework filled with generic text. The tool didn’t malfunction. It merely executed my directive as given.

In subsequent prompts, when I requested quotes with specific outcomes or deadlines achieved, the language improved noticeably to resemble something a client might actually say.

Regardless, one constant remains: I fabricated each name, title, and quote. For any legitimate business, they would require replacement with authentic customer testimonials before going live.

Thus, the layout offered value, while the text did not.

Bear that distinction in mind for the additional eight prompts. In nearly all cases, the layout was sound, but the quality of text varied.

Four Sections Where I Provided an Opinion

This is where the tool truly excelled, and it took me a while to appreciate this aspect.

A Letter From the Founder

Prompt: Three unpopular opinions from a wedding photographer about wedding photography. One sentence each, stated as claims not questions.

In hindsight, this was by far the best output of the day.

I received a concise label stating, “A LETTER FROM THE FOUNDER,” along with the headline “Crafted with care in every small batch,” three paragraphs advocating for the practice rather than merely describing the company, and a signature attributed to Mara Ellis, Founder and Soapmaker.

This was one of only two instances (out of nine) that I would have happily used as it was delivered.

Three Unpopular Opinions

Prompt: Three unpopular opinions from a wedding photographer about wedding photography. One sentence each, stated as claims not questions.

“A streamlined wedding schedule produces more authentic photographs than a fully packed itinerary of formalities.”

This is a valid argument for a photographer, which others may counter. That’s what distinguishes an opinion from a vague statement like “the entire experience is considered.”

The prompt even specified: “stated as claims, not questions.”

Who I Am Not for

Prompt: Create a two-column section for a freelance copywriter: who I am a good fit for, and who I am not. Four honest points each. The second column should explicitly state whom I am not for.

The headline crafted by the AI, “A good fit outweighs a full calendar,” exceeded the expectations of the initial prompt.

Every point in the black card initiated with “”: misleading claims, hurried projects, and endless revisions. The contrasting black card against the lighter one was its concept, not mine.

It also provided three points on each side, despite my request for four. Not a significant flaw, yet noticeable.

A Note About Mondays

Prompt: I requested a brief note from a restaurant detailing why it closes on Mondays.

The generated output included two paragraphs: one explaining the reasoning behind the closure and another apologising for any inconvenience. This was signed, “With gratitude, Chef Marco.”

Typically, a section like this is just a quick line of small print. When you provide a reason, it transforms into something that people might actually read.

The actual triumph across these sections was: I provided the objective, and the AI effectively managed the structure.

Two Sections Where I Only Gave It Facts

However, not every section must have a perspective. At times, what you require is information that is well-organised. The following two prompts tested the generator’s ability to achieve just that.

Pricing

Prompt: Generate a pricing section for an online course. Three plans: Self-Paced $149, Guided Cohort $349, Mentorship $699. Guided Cohort is the most popular.

This section was ready for immediate use.

It presented three columns, highlighted the middle plan in the theme’s accent colour with a “Popular” label, provided a self-generated headline, and featured one benefit line for each plan that built on the previous one.

All I did was provide the prompt and the pricing details, and the AI took care of the rest.

A Size Guide

Generate a shirt size chart displaying chest, sleeve, and length measurements for five sizes.

Five sizes were provided. Three rows were returned.

M, L, and XL grouped into a single row instead of being separated. The generator had all the available data, but I wasn’t specific enough about the intended arrangement.

In hindsight, I should have requested each size be given its own row, with size and measurements displayed together.

This illustrates the importance of context. The more specific you are regarding how you want the information organised, the less the generator has to assume.

Two Heroes, Two Distinct Lessons

The Dentist

Prompt: Create a hero section for Bright Lane Dental, a family dentist in Bristol for anxious patients. Include evening and weekend appointments. Include buttons: Book a Consultation and View Our Prices.

The layout returned precisely as I had instructed: a hero section with a headline, supporting text, two buttons, and an image.

It effectively incorporated the key details from the prompt—Bristol, the family dentist positioning, catering to nervous patients, evening and weekend appointments, and both desired buttons were included.

The main issue arose with the image. I described a dentist for anxious patients, yet it chose a close-up of dental instruments in a patient’s mouth. While it was relevant to the dental theme, it lacked the reassuring context I aimed for.

A more calming image, such as a dentist conversing with a patient, would have been a superior fit.

The Plumber

Prompt: A hero for an emergency plumber in Bristol. Include a headline, the phone number as the central element, and a note stating services available 24 hours. Image on the right featuring a van parked in a residential area at night.

Share this content: