Hook&Hold

The short versionPeople are getting answers without clicking. If AI doesn't mention your brand in its answer, you're invisible. Generative Engine Optimization (GEO) is how you get mentioned.

The click is shrinking

In 2025, Pew Research tracked the real browsing of 900 US adults. Here's what happened when Google showed an AI summary at the top:

And 26% of searches with a summary ended the session right there, versus 16% without one.

Translation: the answer box is eating the click. And it's not just Google. ChatGPT has 800M+ weekly users, and eMarketer projects 31.3% of the US population will use AI search in 2026 (eMarketer).

So is SEO dead?

No. Google still handles roughly 417 billion searches a month, compared to ChatGPT's roughly 72 billion messages (eMarketer). Search is still huge.

But the goal changed. Old SEO asked, "Do I rank?" GEO asks, "When someone asks AI about my category, does it say my name?"

And here's the twist that should get your attention: the AI traffic you do get is gold. The Washington Post found subscribers coming from AI platforms converted at 4–5x the rate of regular search visitors. Fewer clicks, but they show up ready to buy.

How AI decides who to mention

Nobody outside the AI companies knows the exact recipe. But the patterns are clear:

1. Third-party talk beats your own website. Reddit, LinkedIn and YouTube are among the sources AI quotes most. What others say about you matters more than what you say about yourself.
2. Mentions beat backlinks. In old SEO, links were the currency. In GEO, being named in trusted places counts more.
3. Answers up front win. AI pulls content that answers the question in the first sentence or two. Long warm-ups get skipped.
4. Fresh beats stale. Updated content gets picked over old pages.
5. It's a moving target. 40–60% of the sources AI cites change month to month. You can't "rank #1" and coast.

The GEO playbook: 6 moves

1. Run a "prompt audit." Ask ChatGPT, Gemini, Perplexity and Claude the top 10 questions your customers ask. Log who gets mentioned. Repeat monthly, since the answers shift.
2. Put the answer first. Rewrite your key pages so the first two sentences answer the question directly. Save the story for later.
3. Build an FAQ that sounds like real people. Use the actual questions customers ask, word for word.
4. Get talked about off your site. Encourage reviews. Show up helpfully on Reddit and in industry forums. Pitch guest posts and podcasts.
5. Make your facts easy to grab. Clear pricing, specs, locations and hours. AI can't recommend what it can't understand.
6. Update on a schedule. Refresh your top 10 pages every quarter with new data and a new date.

How to measure it

You can't see prompt volumes or exactly why AI picked you. But you can track:

• How often you're cited across AI tools
• Your share of voice versus competitors
• Referral traffic from AI platforms
• Whether mentions are positive or negative

Sources: Pew Research via 9to5Google · eMarketer — FAQ on GEO and AEO

The short versionAI agents are starting to research, compare and even buy things for people. That means your product page now has two audiences: humans and machines. Most brands are only writing for one.

This is already happening

It sounds like sci-fi, but the numbers say otherwise. According to an Elogic roundup of Adobe, Salesforce and OpenAI data:

• ChatGPT handles about 50 million shopping questions a day.
• AI-driven traffic to US retail sites grew 393% year over year in Q1 2026 (Adobe).
• In March 2026, AI-referred shoppers converted 42% better than everyone else. A year earlier, they converted 38% worse.
• Salesforce estimates AI agents drove about 20% of global orders in the 2025 holiday season.

That conversion flip is the headline. A year ago AI shoppers were window-shopping. Now they show up ready to buy.

The plumbing is being built right now

The big players are racing to set the rules for how AI agents pay for things:

You don't need to understand the tech. Just know this: when Visa, Mastercard, Google, OpenAI and Shopify all build the same thing at once, it's not a fad.

How big could it get?

Forecasts for 2030 vary, but they all point the same way:

• McKinsey: $3–5 trillion in agent-driven commerce worldwide
• Morgan Stanley: 10–20% of US e-commerce
• J.P. Morgan: up to 25% of US online sales

Even the cautious guess is a massive shift in how people buy.

The catch: people don't fully trust it yet

Here's the reality check. About 62% of consumers use AI to compare products, but only 23–30% trust it to finish a purchase on its own. People love AI as a research assistant. They're not ready to hand it their wallet.

That gap is your opportunity. The brands that feel safest to buy from, with clear prices, easy returns and strong reviews, will be the ones agents and humans both pick.

What "marketing to a robot" actually means

An AI agent doesn't care about your clever headline or your hero video. It cares about clean, complete, consistent facts. Think of it like a very picky buyer who reads every spec and trusts nothing vague.

One stat to remember: pages with structured data (code that labels your price, rating, stock, etc.) were cited about 3.1x more often in Google's AI Overviews.

6 moves to make your brand agent-ready

1. Fill out every product field. Price, size, materials, shipping time, return policy. Blank fields make agents skip you.
2. Add structured data (schema markup). Your web developer or a plugin can do this in an afternoon. It's the single highest-leverage fix.
3. Keep facts the same everywhere. If your site says $49 and Amazon says $52, agents lose trust.
4. Make your policies dead simple. "Free returns within 30 days" beats three paragraphs of legal text.
5. Stack up reviews. Agents lean on social proof just like humans do.
6. Fix your tracking. About 70% of AI referral visits may not show up correctly in a standard GA4 setup. Set up custom channel groups for AI sources so you can see what's working.

Source: Elogic — ChatGPT Commerce & Agentic Shopping Statistics 2026 (compiling Adobe Analytics, Salesforce, OpenAI, McKinsey, Morgan Stanley and J.P. Morgan data)

The short versionWhen AI answers a question, it often quotes Reddit. That makes a bunch of strangers in a comment thread one of the most powerful marketing channels on earth. You can't buy your way in. You have to earn it.

The internet's most-quoted forum

When researchers look at which websites AI tools quote most, Reddit keeps landing at or near the top (QuickSEO):

One cross-platform study found Reddit cited in 40.1% of cases across the major engines, ahead of Wikipedia (26.3%) and YouTube (23.5%).

This didn't happen by accident. Google signed a deal with Reddit in February 2024, reported at about $60 million a year, for access to its content. OpenAI signed its own deal in May 2024.

Why AI loves Reddit

Think about how you shop. You read the brand's website, then you search "[product] reddit" to find out what real people actually think. AI does the same thing, at scale.

Reddit has what brand websites don't: unfiltered human experience. "I've used these boots for two years on a ranch and the soles finally cracked" is worth more than any product description.

Warning: it's volatile

Don't bet the farm on one number. According to Semrush data, Reddit appeared in nearly 60% of AI responses in early August 2025, then dropped to around 10% by mid-September 2025 (The State of Brand). The AI companies keep adjusting their models.

The lesson: Reddit is a major input, but the weights shift. Play the long game.

The trap: faking it

When marketers smell opportunity, some try to game it. Agencies now sell packages of Reddit "mentions," around $2,500 a month for 35–100 posts. This is a terrible idea, for three reasons:

1. It's illegal if undisclosed. The FTC's rule on fake reviews and testimonials has been enforceable since October 2024. Penalties can reach $53,088 per violation.
2. Moderators will catch you. Subreddits have cracked down hard. Some limit self-promotion to once every 60 days. Others ban it outright.
3. It backfires publicly. When one jewelry brand's job posting for Reddit marketing leaked, moderators warned members and the story landed in the Wall Street Journal. That's the opposite of what they wanted.

Fake mentions also hurt the thing that makes Reddit valuable: authenticity. AI models are getting better at spotting it.

The honest Reddit playbook

1. Listen first. Search Reddit for your brand, your competitors and your category. Read what people complain about. That's free market research.
2. Join as a real person. Use your own account. Say who you are: "I work at X, so I'm biased, but..."
3. Follow the 90/10 rule. Nine helpful comments for every one that mentions your product.
4. Answer the question first. Be the most useful reply in the thread, even when the answer isn't you.
5. Fix what people complain about. The best way to change Reddit's opinion of you is to change your product.
6. Track it over time. Check monthly what AI says about you and which threads it's pulling from.

Sources: QuickSEO — How Reddit affects AI visibility in 2026 · The State of Brand — Reddit AI citations and brand risk · eMarketer — Creators, AI search and social

The short versionBig social feeds are crowded, noisy and run by algorithms you don't control. Smaller, tight-knit communities are where trust lives now. A room of 500 true fans can beat a feed of 50,000 strangers.

The feed is getting tired

Remember when you could post on Facebook and actually reach your followers? Those days are gone. Algorithms decide who sees what, and AI is flooding feeds with more content than ever.

So people are moving. Not off the internet, but into smaller rooms. Discord servers. Substack newsletters. Private Facebook groups. Group chats. Niche subreddits.

At Social Media Week, Discord's VP of sales put it simply: "Community is the product" (Adweek).

What the data says

Kantar named micro-communities one of its top marketing trends for 2026. Two numbers stand out:

• Nearly 40% of consumers trust recommendations from micro-communities as much as advice from people they know personally.
• In China, brands using knowledge-sharing community platforms saw 25% higher marketing ROI. (That market is ahead of the US on this, which makes it a useful preview.)

eMarketer reports the same shift: brands are moving focus from saturated main feeds to niche spaces on Discord, Substack and Medium, where an engaged audience is easier to manage.

Why small wins

Think of it like this. A billboard on the highway reaches thousands of people who don't care. A recommendation at the local feed store reaches ten people who are ready to buy.

Micro-communities work because of three things:

1. Shared identity. Members aren't just "users." They're runners, home bakers, horse breeders or golf nerds. They talk the same language.
2. Peer trust. A tip from someone who's "one of us" beats any ad.
3. Real conversation. People ask questions, share wins and complain honestly. That's both marketing and market research.

The mistakes brands make

• Showing up to sell. Walking into a community with a sales pitch is like walking into a party and handing out business cards. People leave.
• Chasing size. A community of 300 active people beats 30,000 who never post.
• Going quiet. Communities die without a regular rhythm. If you start one, show up every week.
• Controlling everything. The best communities belong to members, not the brand.

The micro-community playbook

1. Find where your people already gather. Before building anything, join the groups your customers are already in. Listen for a month.
2. Give more than you take. Teach. Answer questions. Share behind-the-scenes info no one else has.
3. Start small and specific. "Marketing for small businesses" is too broad. "Marketing for Oklahoma ranch and farm businesses" is a community.
4. Create rituals. A weekly Q&A, a monthly challenge, a "win of the week" thread. Rhythm keeps people coming back.
5. Spotlight members. Make your members the stars. People stick around when they feel seen.
6. Measure depth, not reach. Track active members, replies, repeat buyers and referrals, not just member count.

Sources: Kantar — Marketing Trends 2026 · Adweek — Discord, Substack say niche communities are the future · eMarketer — Creators, AI search and social

The short versionBrands are pouring billions into creators, but most can't prove it's working. The party isn't over. The bouncer just started checking IDs. Creators who can prove results win. Everyone else gets cut.

The money is real

Creator marketing is one of the fastest-growing places ad dollars go. According to the IAB:

• US creator ad spend hit $29.5 billion in 2024, more than double 2021's $13.9 billion.
• It was projected to reach $37 billion in 2025, growing 26%, about 4x faster than the media industry overall (5.7%).
• 48% of advertisers now call creators a "must-buy," ranking only behind social media and paid search.

And Kantar found 61% of marketers plan to spend more on creators in 2026.

The problem nobody wants to say out loud

Here's the awkward part. Most of that money is hard to track.

• 79% of enterprise marketers say they struggle to measure influencer ROI (Influencer Marketing Hub).
• 48% say attribution is their biggest gap.
• Only 27% of creator content ties strongly to the brand that paid for it (Kantar).

Read that last one again. Nearly three out of four sponsored posts leave people unsure who the sponsor was. That's a lot of money spent making creators famous instead of brands.

Why it's happening

1. One-off posts. A single sponsored video is a sugar rush. People forget it in a day.
2. Chasing big follower counts. Followers don't equal influence. Many accounts have huge numbers and little trust.
3. Over-scripting. When brands force creators to read stiff talking points, audiences tune out.
4. Measuring the wrong things. Likes and views feel good but don't pay the bills.

What's changing

Smart brands are already adjusting:

• Going smaller. About 53% plan to expand micro-creator deals and 51% nano-creator deals. Smaller creators often have tighter, more trusting audiences.
• Going longer. Kantar recommends moving from one-off posts to "long-term creative platforms," meaning ongoing partnerships with the same creators.
• Using guardrails, not scripts. Kantar's advice: set clear guardrails, "then let creators do what they do best."
• Demanding proof. As eMarketer put it: creators who can prove they influence real humans will become premium inventory.

How to run creator deals that actually pay

1. Pick fit over fame. A creator with 8,000 followers who are exactly your customer beats 800,000 random ones.
2. Brand it early. Show or say your brand in the first few seconds. Make it part of the story, not an ad stuck on the end.
3. Track everything. Give each creator a unique link and discount code. No code, no proof.
4. Commit to a series. Plan 3–6 posts over a few months instead of one big splash. Repetition builds memory.
5. Get usage rights. Pay a little extra to run the creator's content as your own ads. Often the best-performing ads you'll have.
6. Judge by business results. Sales, sign-ups, leads and cost per customer. Views are a bonus, not the goal.

Sources: IAB — Creator Economy Ad Spend 2025 · Kantar — Marketing Trends 2026 · Influencer Marketing Hub — Influencer marketing statistics · eMarketer — Creators, AI search and social

The short versionAI made content cheap. When everyone can make 100 ads in an hour, making stuff stops being the skill. Knowing which one is good becomes the skill. Taste is the new superpower.

Everyone's using AI. Customers noticed.

Canva's State of Marketing and AI Report (May 2026, 1,415 marketing leaders and 3,547 consumers in seven countries) shows a huge gap:

Marketers are going all-in on AI. Customers are quietly asking for a human in the room.

The "slop" backlash is real

"AI slop" is the internet's name for low-effort, mass-produced AI content. And people are tired of it.

• Mentions of "AI slop" rose about ninefold in media monitoring (Canva).
• Consumers preferring AI-generated creator content over traditional creator content fell from 60% in 2023 to 26% (Digiday, citing Billion Dollar Boy).
• Some big brands, including Polaroid and Heineken, have even run "human-made" campaigns as a selling point (Outlier Report).

As one agency CMO told Digiday: "AI can't replicate the messiness of human creativity. We crave that now, we crave imperfection."

Wait, so should we stop using AI?

No. That's the wrong lesson. AI is an incredible tool for research, drafts, editing and testing ideas fast. Nearly every top marketer uses it.

The right lesson: AI is the kitchen, not the chef. It can make a hundred dishes. Someone still has to taste them and decide which one goes out to the customer.

That someone is you.

What "taste" actually means

Taste sounds fuzzy, but it's a real, learnable skill. It's the ability to quickly tell:

• Clear vs. confusing. Would a busy person get it in 3 seconds?
• Specific vs. generic. Does it sound like this brand, or like everyone?
• True vs. fake. Does it feel like a real person said it?
• Surprising vs. expected. Would anyone stop scrolling for it?

AI is great at "expected." Humans win at "specific," "true" and "surprising."

How to build your taste (on purpose)

1. Build a swipe file. Save every ad, email or post that makes you stop. Once a month, look for patterns in what you saved.
2. Study the greats. Read old ads from legendary copywriters. Good ideas age well. Gimmicks don't.
3. Generate wide, pick narrow. Use AI to make 20 versions. Then cut to the best 2. Picking is the skill you're training.
4. Explain why. For every choice, write one sentence on why it beats the others. If you can't explain it, you don't know it yet.
5. Add the human fingerprint. A real story, a real photo, a real customer quote, a little imperfection. That's what AI can't fake.
6. Test it. Taste is a guess. Data checks the guess. Run A/B tests and see if your instincts hold up.

Sources: The Next Web — Canva State of Marketing and AI Report 2026 · Digiday — Authenticity and 'messiness' in high demand · Outlier Report — Anti-AI is the premium · Marketer Milk — 8 marketing trends in 2026

The short versionBig dreams like a house feel out of reach for a lot of people. So they're buying small joys instead: a fancy coffee, a candle, a concert ticket. Marketers call it "treatonomics," and it's one of the most important consumer moods of the year.

Meet the "little treat"

You've seen it on social media: "I had a rough day, so I got a little treat." It's a joke, but it's also a real spending pattern.

A SurveyMonkey study of 2,038 US adults (February 2026) found:

• 62% of Americans buy themselves a small treat at least once a month.
• 43% do it daily or weekly.
• 52% keep it to $25 or less per treat.

Kantar calls it "the lipstick effect on steroids." The classic lipstick effect says that when times are tough, people skip big purchases but still buy small luxuries. Kantar also found 36% of consumers are willing to take on short-term debt for things they enjoy.

What people are treating themselves to

Food wins by a mile. But notice hobbies and experiences right behind. People aren't just buying things. They're buying feelings.

Why they do it

The reasons are the real gold for marketers:

1. To stay motivated toward bigger goals (35%)
2. To cope with stressful days or bad news (30%)
3. To celebrate personal wins (27%)

Gen Z and Millennials lead on all three. For younger buyers, a little treat is part reward, part therapy and part celebration.

The insight: sell the moment, not the product

Kantar's advice to brands is to create joy through everyday moments instead of big, aspirational dreams. Translation:

• Old message: "Someday you'll have the life you want."
• New message: "You earned this today."

That's a huge shift. Your product isn't just a coffee or a candle. It's a reward for getting through Tuesday.

A responsible word of caution

Here's the part most trend articles skip. The same study found frequent treaters are nearly twice as likely to say treat spending hurts their financial goals (30% vs. 16%). Millennials feel that pressure most.

Good marketers don't exploit stress. The brands that last will make treats feel good, not guilty. Think fair prices, honest value and no pressure tactics. Trust is the real long-term treat.

The treatonomics playbook

1. Create a "treat-size" option. A mini version, a sample pack or a single-serve product under $25.
2. Name the moment. "Friday reward," "Survived Monday," "Small win." Tie your product to a feeling and a time.
3. Make it easy to grab. Treats are impulse buys. Checkout should take seconds.
4. Celebrate your customers' wins. Birthday perks, milestone rewards, "you hit your goal" emails.
5. Sell experiences, too. A workshop, a tasting, a behind-the-scenes visit. Memories are treats people love to share.
6. Keep it honest. Never guilt or pressure. Make people feel good about the purchase after they buy it.

Sources: SurveyMonkey — 2026 Treatonomics Report · PPC Land — Kantar forecasts ten marketing trends for 2026 · Kantar — Marketing Trends 2026