When this article was first published in 2023, generative AI looked like an opportunity to make almost every part of SEO faster. Some of that proved true. AI is useful for sorting research, maintaining content, and helping people work with data.
The more ambitious promises did not age as well. AI does not reliably predict algorithm updates, automatic content does not earn rankings simply because it is optimized, and voice search did not require a separate content strategy for most small businesses.
Here is a more useful assessment.
What held up
Faster analysis
AI can categorize search queries, summarize customer feedback, identify repeated questions, and help clean analytics exports. That gives a marketer more time to interpret the evidence and decide what to do.
It works best when you provide the source material. Asking a model to invent keyword demand, audience insights, or performance data is not research.
Better content maintenance
Reviewing an existing site is a strong use case. AI can help locate:
- pages that cover the same topic
- outdated references and broken links
- vague statements that need evidence
- titles and descriptions that say the same thing
- sections that repeat the introduction without adding information
An editor still needs to verify the findings and choose what to update, merge, or remove.
Help with repetitive production work
Drafting alt text for review, converting notes into a brief, writing regular expressions, and generating schema markup from verified page data can all save time. These are bounded tasks with an output a person can check.
What was overhyped
“Predictive SEO”
Historical data can reveal seasonality and trends. It cannot tell you with confidence what a search engine's next update will reward. Building a strategy around predictions about an undisclosed ranking system is speculation, even when a model presents it fluently.
Publishing at scale
Generating more articles is easy; producing information that customers trust is not. Pages assembled from the same public sources tend to repeat what already exists. Google's spam policies also warn against scaled content created mainly to manipulate rankings, regardless of whether automation or people produced it.
Optimizing for every search format
Voice, visual, and AI-assisted search are real, but most businesses do not need a separate page for each format. The durable work is less exciting:
- answer the question directly
- use descriptive text and useful images
- make the site accessible and crawlable
- cite original sources
- keep important information current
- add genuine experience that another site cannot copy
A sensible way to adopt AI
Choose one time-consuming task with a result you can verify. For example, use AI to group six months of search queries by customer need. Check a sample against the raw data, then use the groups to improve pages that already serve those customers.
Keep records of the inputs, prompts, and editorial checks for repeatable work. Do not paste confidential customer or business data into a tool unless its terms and data controls are appropriate.
AI can be a useful assistant for SEO. The opportunity is operational—not a shortcut to authority, originality, or trust.
For a detailed working method, continue with Using AI for SEO Without Publishing Generic Content. If your immediate need is a stronger website or local presence, see Amber Tribe's website and Google Business services.