How (and Why) to Auto Generate Website Content

If you’re a left-brain thinker, it might prove challenging to produce compelling content. As Marketing AI Institute’s resident math geek Keith Moehring admits, “I hate writing. Always have.”

For left-brained content marketers, the ability to auto-generate website content is a temptation too seductive to resist.

Ambiguity and nuance — two strengths content creators can’t function without — frustrate left-brainers. However, Moehring has a point. Some types of content do indeed follow a formula of sorts.

Great writing (or even great video content production) rests on underlying structures and formulas. Take the hero’s journey, for instance — the stuff of practically every movie Hollywood produces — and a valuable formula (with a few tweaks) for marketing content.

Image via Grand Valley State University

Casting your prospective customers as the hero and positioning your advice or products as the sword that helps them slay their pain points is a genius move. It helps prospects invest themselves emotionally in the story and follow through until the end.

But therein lies the rub. It takes a good grasp of nuance to draw an audience into the story. So far, only human content writers have the creative chops to engage their audience to the end of each piece.

Where Does Auto-Generated Content Fit into My Content Operations?

Auto-generated content can indeed play a supporting role in your content operations. With close human supervision, it can save content marketing teams a lot of time by automating routine tasks.

Create Routine Documents with Predictable Structure

Natural language generation (NLG) software has the capacity to fill in the blanks on routine documents with similar structures, such as earnings or content analytics reports, performance reports, feature lists on product pages, and bio pages on your website. Here’s how:

  • First, create your templates: Sketch out the skeleton of various types of content that can utilize structured data (think anything you can put in spreadsheet format).
  • Next, choose where your variables will go in the content: These are the “blanks” where the NLG program will insert relevant data points from the spreadsheet. Use the names of each data point (the headings on each spreadsheet column) as your variables.
  • Then, write your conditional statements: These if-then statements will determine which content the NLG content auto-generator will print out on your copy. For example, if you’re looking to list negative departmental output in an internal report, you would write, “If departmental output < 0, then print.”
  • Finally, choose synonyms that the NLG can insert for variety: For instance, in the example below, the original word is “below.” Synonyms could include “under,” “less than,” or “lower than.”

Image via Marketing AI Institute

Auto-Generate Product Descriptions

If your company manufactures or sells products at a large scale, AI can generate lists of features much faster than human writers.

However, have a human writer and editor review and tweak the content before you click “Publish.” That’s especially true if you have a descriptive section in addition to bulleted lists of features. Again, even the most sophisticated AI content generators can’t grasp the nuanced copy that entices customers to buy the product.

If you’ve ever used Grammarly to check for typos and grammatical errors, you’re well aware of the limitations of AI-generated content.

It can get a little awkward. For example, in a descriptive paragraph I once wrote, Grammarly suggested that I substitute the word “reckon” for “measure” for a description of a device’s function.

“Reckon”? Only in the Deep South. And only — even there — in casual conversation.

And then there was that wristwatch description where Grammarly needled me to substitute “observations” for “watches.” No nuance, no grasp of context when a word has ambiguous meanings. So always check your AI content auto-generator’s work, lest you make an embarrassing mistake that hurts your company’s business.

So, don’t fire your content creation staff just yet. You need them to add the compelling nuances that no machine can generate. But you can save them some time so that they can spend more time on creative tasks.

Auto-Generate Topics, Headlines, Outlines, and Drafts with Language-Capable AI

Language-capable AI programs like BERT and GPT-3 can combine texts about a specific topic to come up with a passable first draft of blog posts or static web content. However, as the Search Engine Journal’s Shelley Walsh points out, AI content creators haven’t advanced far enough to create data-driven content, do extensive research, or generate the kinds of innovative ideas that drive thought leadership pieces.

For that reason, auto-generated content works better as a topic idea generator, outline creator, or, at best, generating a rough draft for blog posts. Once your human writers have created a fleshed-out post from the draft, however, your AI content generator can digest that content and come up with a decent meta description.

Language-capable AI can also generate keyword-rich headlines. However, like other auto-generated content, double-check them for awkward word order or fuzzy meaning.

And when writer’s block hits, having an AI content generator around can help you over the hump by suggesting topic ideas — usually the sticking point with most marketing content writers.

Auto-Update Website Content

Older posts or static website content often contain outdated statistics or facts. During your next content audit, set the time-sensitive portions of your content as variables that update automatically, populating the content with current data. Auto-updating your content keeps your facts fresh, likely generating more shares and engagement.

Again, be sure to check the changes to make sure that everything is grammatically correct and makes sense.

If You Use Auto-Generated Website Content, Follow Google’s Guidelines

Google has strict guidelines for auto-generated content. Make sure that any AI-created content you publish online conforms to their standards.

Google will penalize auto-generated content that:

  • is of such low quality that it makes little sense but features search keywords
  • is auto-translated (like with Google Translate) without a native-speaking or fluent editor’s touch
  • results from automated processes like Markov chains
  • comes from scraping search results or RSS and Atom feeds
  • comes from obfuscation techniques or automated synonymizing
  • is the product of combining content from various web pages without adding value

Again, it’s superb content that puts your brand at the top of searches. Even more importantly, it’s what helps transform casual browsers into loyal customers.

Creating that kind of content takes organizational efficiency from top to bottom across the enterprise. From content collaboration with subject matter experts to integrations with third-party software, DivvyHQ can manage the entire content production process from the first germ of an idea to publication and analyzing your content’s impact on its target audience.

It’s the content marketing platform of the future. And you can try it free for 14 days with no obligation. Start your free trial today!