Solving the Google AIO vs. Discover conflict
For decades, the “inverted pyramid” was a standard in journalism: put the most important facts at the top and let the details trickle down. Then came the SEO era, where we learned to “bury the lead” just enough to keep people scrolling, stretching engagement and ensuring readers saw more display ads.
Now, Google is forcing us back to the pyramid with a cruel twist that has forced publishers into corner.
If you want your content to surface in Google’s AI Overviews (AIO) to build brand authority and drive targeted traffic, you have to be straightforward. The AI models are looking for “fact density” and “extractability.” They want the key information in the first 100 characters. If you don’t give it to them, you won’t get the citation.
So here’s the conflict…
By giving the answer away upfront to win the AI slot, you are hurting your chances of being surfaced in Google Discover, which, for many of us, has become a much larger source of referral traffic than traditional search. For publishers aiming for big numbers, Discover is infinitely more important that AIO at this point – but that doesn’t mean you can’t have it both ways. We’ll get to that in a minute.
In a nutshell:
AI Overview rewards “Direct Answer” content. If a user asks “Who won the Rams game?” and you say “The Los Angeles Rams beat the Dallas Cowboys 21-14 on Sunday” in the first paragraph, Google’s AI is more likely to scrape that and feature you as a source.
Google Discover rewards Engagement Time and serves users content based on their interest, browsing history, location, and other factors. If a user clicks your link from Discover, finds the Rams score in three seconds, and then bounces, Google’s algorithm marks that as a “low-quality” interaction and probably won’t surface it to many users.
How to Target AIO
There is intrinsic value in being a reputable, cited source in Google across all of its products. Especially if your targeting specific search terms and questions.
To rank in the AIO box, our old pal, Associated Press Style, is a good starting point. Deliver the critical information right away, then treat the rest of the content almost as if you’re a database, even incorporating bulleted information. Forgo flair.
The 2026 playbook for “Answer Engine Optimization” is this:
The 60-Word Rule: Use H2 headings that are literal questions, immediately followed by a 40–60 word “definition block.” No fluff.
For example: “What time does the Rose Parade start?”
“The 2026 Rose Parade in Pasadena starts at 8 a.m Pacific Time. The starting point is at…” and so on.
The “Fact Density” Requirement: AI models ignore vague prose. They don’t care about your colorful writing. They want information and data. If you lead with “A survey of 500 Gen Xers found 62% prefer Coke to Pepsi” you are 4x more likely to be a cited source.
Modular Formatting: Use bulleted lists and tables. Over 75% of AIO responses now favor list-based formatting because it’s easy for the model to strip mine.
The Content “Organ Donor”
When you follow these rules, you aren’t really a publisher anymore. You are breaking your body of work into small, meaty chunks for Google to feed to its users so they don’t have to visit your site. It doesn’t sound like much fun, and it sure doesn’t sound like a recipe for engagement.
By targeting the AI snippet, you are deciding to give up the engagement, or “dwell time,”that Discover prefers.
Recent data shows that organic click-through rates (CTR) can plummet by over 60% when an AI Overview is present.
News publishers are seeing this unfold in real-time with the Discover Core Update in February 2026, which transparently signaled that Google is getting even more selective, rewarding sites that offer deep E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) rather than just quick answers. KTLA and many other publishers, even those with strong E-E-A-T, are feeling the impact acutely and seeing it play out in our analytics.
If you provide the “hero” information upfront, the user has no reason to click. If they do click, they don’t stay. You’re being asked to choose between being a source for a robot or a destination for a human.
Splitting the Difference: Contextual Layering
So, how do you win on both front? We’re seeing a shift toward “Contextual Layering.” Instead of just giving the answer, publishers are beginning to lead with a “Yes, and…” approach to highlight depth.
You provide the immediate payoff to satisfy the AI’s needs, but then pivot to unique value. This includes proprietary data, first-person experiences and analysis to keep humans engaged. Interviews and quotes are useful for this. You satisfy the robot’s need for facts, but you give the human reader a reason to stick around.
To see how this works in the real world, let’s look at the January 2026 California housing data, which is provided by the California Association of Realtors (yes, they own the AIO box) . Here is how you would use Contextual Layering to try to win on both fronts:
Layer 1: Targeting AI
What was the California median home price in January 2026?
According to the California Association of Realtors (C.A.R.), the statewide median home price in January 2026 was $823,180. This was a 3.2% decline from December 2025 and a slight drop from the $839,130 recorded in January 2025, marking a 23-month price low for the state.
Layer 2: Targeting Human Engagement
While California’s median home price was flat to down, that wasn’t the case in every region of the state.
San Francisco and San Diego saw modest gains last year. Joe Realtor of Real Estate Inc. told us that his open houses are often packed, and competitive offers come within hours, not days or weeks. “A lot of home buyers have decided they can’t wait for mortgage rates to come down,” Joe said.
The Bottom Line
Google holds a lot of sway and publishers need to constantly adapt to its core updates and changing algorithms. One product wants publishers to be an answer engine that renders your website unnecessary. The other wants you to be a content feed that delivers engagement. Yes, Google News and Google Search are still very much relevant, and these approaches have the potential to work there as well. Authority, trust and a strong SEO gameplan can deliver on all fronts.
As publishers, we are caught in the middle. Those who find a way to serve both of Google’s needs, giving the AI what it wants while keeping the soul of the story for the reader, will be the ones who survive and thrive.
