More than 300 marketing pros told us whether their worst fears or wildest dreams have come to fruition since the launch of Google’s AI Overviews (formerly SGE) in May 2024.

The much-hyped and much-criticized event, which Google immediately scaled back after AI Overviews (AIO) served users with incorrect and potentially harmful info, marked the wide launch of AI-generated search on the world’s biggest search engine.

Now that Google has patched up its early missteps and SEOs have had time to observe traffic changes, we conducted a survey to determine whether the white-knuckled AI anxiety is backed by any evidence.

We’ve got the numbers on AIO’s effects on traffic, what SEO changes to look for in the coming year, and pro tips from SEO experts on how to adapt and optimize your web and content strategies.

Table of Contents

During its beta phase, Google’s AI-powered search was called SGE, or Search Generative Experience. It’s now more commonly known as AI Overviews, and it’s Google’s response to AI-powered search engines like Perplexity, Brave, and Microsoft’s Copilot.

It runs on Google’s AI model called Gemini, formerly Google Bard.

Regular ol’ Google Search still uses a three-stage process, though Google notes that not every page will make it through each stage. Google Search first crawls web pages, downloading text, images, and videos; it then indexes that information and stores it in a database; finally, it serves search results to the user.

AI Overviews (AIO) integrates generative AI capabilities with Google Search results. Using natural language processing (NLP), AIO can effectively repackage search results as answers to your questions, not just a page of links.

Not every query will trigger an AIO response — a Google spokesperson reiterated to me that AIO is designed to appear when it’s helpful beyond the usual SERPs. And since web content changes and evolves, AIO is also dynamic.

If you’re not seeing any AIO, make sure you’re logged in to your Google account, and check Google’s growing list of countries where it’s available.

As of October 28, 2024, Google added more than 100 countries and territories to AIO list, including the U.K, Japan, Canada, Mexico, and Brazil. If you’re in Spain, France, Germany, or Ireland, you’ll still see only the traditional SERPs.

Here’s an example of how AIO differs from the usual SERPs. If I search “world without gravity,” AIO answers the questions it thinks I have, with source links for each point:

Image Source

Google unfortunately couldn’t have anticipated the amount of anxiety this search just caused me, but in terms of search intent, it’s done a pretty good job.

It found a number of popular science websites that have covered the question, which are listed on the right, and it used that information to organize questions and answers in natural language.

The more familiar SERPs still appear below the AIO:

Image Source

Braden Becker, a former HubSpotter who’s currently the global SEO lead at Faire, began with an important caveat about the effects of evolving AI in search, generally speaking: “I will say it depends on the industry. I think that‘s one thing I don’t hear enough of — this is not going to affect everyone equally.”

Some industries may be already feeling the effects, and some may not see any noticeable traffic changes for years.

Amanda Sellers, manager of EN blog strategy at HubSpot, says that “AI’s increasing ability to enforce E-E-A-T standards could push marketers to create content that’s richer, more nuanced and specific, and better accounts for search intent.”

Aleyda Solís, international SEO consultant and founder at Orainti, adds that Google is in a challenging position right now.

It has to make some choices about how it proceeds in the AI search landscape, and it has to match the quality, speed, and user satisfaction of these new players in the AI search game. The paradigm shift could, for the first time ever, pose a real threat to the search giant.

In other words, if users abandoned Google for other AI-powered search engines, it would obviously be bad for Google — “but it would be equally bad if [Google] tried to compete with ChatGPT or completely change their search interface all of a sudden. They cannot do that because they have skin in the game; they have a lot of adverts to show. So they’re trying to react [to AI searches] accordingly without harming their business model, which is extra challenging.”

Google seems to be stepping up to the challenge. In October 2024, Google announced a handful of updates that suggest a focus on alternate search channels, like video and voice.

In brief, if you use the Google app on mobile, you may see these updates:

• Generative AI now works with Google Lens, so you can point your camera, ask a question, and get an AIO response.
• Video understanding capabilities and voice questions: You can ask a question while taking a video or a photo to get an AIO response.
• Shop what you see: Snap a photo of a product you like, and Google’s AI models and the Google Shopping Graph will identify the item.
• Circle to Search: Available on Android devices, this lets you identify songs without switching apps.
• Search results pages organized with AI: Google promises that this will return “relevant results organized just for you.”

So how will AI Overviews change in 2025?

The better question might be, “How will AI Overviews change marketing in 2025?”

Most of the SEOs I’ve spoken with this year have noted a few patterns emerging from the last year of algorithm updates, and they generally see a need for content marketers to make these considerations:

Renew your focus on search intent and follow-up queries.

Continue writing content for your audience — not search engines — by considering user intent and follow-up queries.

Sellers says, “Google is so big as a distribution channel for primarily text-based content. And that means they are often the driver for best practices on our websites. But Google is emphasizing that we shouldn’t be writing content for Google — we should be writing content for our audience.”

Evaluate where your content is reaching your customers, and whether that best serves your interests as AI searches increase.

If you’re in an industry where people even think your product or service is replaceable with AI — even if they’re incorrect — “you need to pay attention, and you need to adapt,“ says Becker. “You’ll definitely need to shore up your product-market fit positioning on how you use top-of-funnel content,” he says, “to clarify how you‘re different and why you’re still valuable.”

On the other hand, if you’re in “an extremely congested market where it’s likely that folks will research their options a bit differently using AI, tofu content — top-of-funnel content — might not be your biggest lever anymore.”

In those cases, Becker says that “you might consider investing more in middle-of-funnel content, trying to reach folks when they’ve already done the bulk of the research.”

Beckers see this as a welcome change. “I think it ultimately reduces content bloat and forces businesses to take a closer look at who they’re really trying to market to.”

Evaluate the specificity and helpfulness of your content.

At the end of the day, creating as much content as possible for a very general audience isn’t going to help your brand unless you’re considering whether your content is actually helpful.

Aja Frost, senior director of global growth at HubSpot, says to take the “experience” in Google’s E-E-A-T quality rating guidelines seriously. “It does really require explaining why you or the author are uniquely positioned to give advice,” she says.

An author bio is a good starting point, but ideally, those explanations should be woven into the content itself. (And it’s why I interview experts for my own articles.)