Can AI Build an Effective Website?

You see what I did in the title there? I threw in a qualifier, effective. Because while AI can certainly build a website, the process is not hugely different from the template driven, automated website builders that have been around for decades. The limitations of those systems are quickly reached when site owners need something specialized, either visually or functionally.

As a professional builder of websites, I’ll be the first to admit that I’m biased on this subject. But I’m also open to new technologies that can streamline the planning, design and building processes that I’ve developed over the years. After experimenting with several of the AI options that are available in tools that I already use, I find myself using them less and less. This aligns with an emerging sense in the tech community that maybe AI is not really coming for our jobs.

I use the term “AI” but in truth there’s little intelligence involved. As Greg Rosalsky writes in 10 reasons why AI may be overrated:

As the technologist Dirk Hohndel has said, these models are just “autocorrect on steroids.” They are statistical models for prediction based on patterns found in data. Sure, that can have some cool and impressive applications. But “artificial pattern spotter” — or the more traditional “machine learning” moniker — seems like a better description than “artificial intelligence.”

These systems don’t have judgment or reasoning. They have a hard time doing basic things like math. They don’t know right from wrong. They don’t know true from false.

Planning, Design, Implementation

Broadly speaking, these are the steps needed to build an effective website. In the planning phase, questions that may need to be answered include:

• Who are the audiences for the site?
• Is there existing branding and marketing that should be included?
• What information will help site visitors make a buying decision?
• What actions are possible on the site?
• How will the effectiveness of the site be measured?

Answers to questions like these are unique to each business and are unlikely to be found in the Large Language Models that power AI.

The design of WordPress websites is controlled by templates (a.k.a. themes), as are the AI website builders. I’ve always been somewhat baffled by how templates are assigned to different business types, since a “lawyer” template can work equally well for a plumber or a donut shop. The choice of template should be driven by the content that was identified in the planning phase. And it must be customizable for unique content. AI may be able to provide options but the best fit will come from a decision made by a human.

After deciding what goes into the website and how it will look, implementation is where the rubber meets the road. The site’s content (text, images, video, etc.) must be easy for the site visitor to find and consume. Any actions that can be taken (e-commerce, site registration, contact forms, etc.) must be configured and tested. In many cases, there is more than one way to go about doing this. Remember that AI can only act based on examples that have been scooped into its model. It may choose the best one, it may not.

Is it helpful?

With any new technology, the effort to use it must be offset by benefits. I can’t speak to the experience of a novice user building a simple website with AI but when I’ve tried to incorporate it into my workflow, it’s often a frustrating experience. For example, not long ago I needed to clean up a number of pages that included lots of non-standard HTML. It seemed like a perfect application for AI, which I thought would know the difference between valid and invalid HTML. After several hours of tinkering with prompts (including a few that almost worked), I gave up. Any technology that is mature enough to use in your business has to work predictably and consistently.

For a simple website that looks and feels like others of its kind, an AI-generated site may work well. However, for something more tailored to your business, a lot of human interaction will be needed. I will close with item number 10 from the article mentioned above:

Reason 10: AI is overrated because humans are underrated.

When I asked [MIT economist Daron] Acemoglu for his top reasons why AI was overrated, he told me something that warmed my heart — a feeling that dumb “artificial intelligence” could never experience.

Acemoglu told me he believed AI is overrated because humans are underrated. “So a lot of people in the industry don’t recognize how versatile, talented, multifaceted human skills and capabilities are,” Acemoglu says. “And once you do that, you tend to overrate machines ahead of humans and underrate the humans.”

Go, Team Human!

Image generated by Adobe Firefly.