At a glance
Who can actually do the work
What it takes to get found on Google and in AI answers
| What it takes | AI tools | SEO agency |
|---|---|---|
| First drafts and outlines | ✓Fast and good | ✓Yes, we use AI too |
| Technical fixes on your website | ✓Can make the fixes | ✓Yes, and we check they worked |
| Rankings and AI visibility tracking | ✓Can track it | ✓Tracked every month |
| Every fact checked and sourced | ~Partly, still needs a human check | ✓Every number verified |
| Original insight from your founders and executives | ✕Can’t interview anyone | ✓Monthly interviews |
| Press mentions and placements | ✕Can only draft a pitch | ✓Lands press, podcast and blog features |
| Analysis and human insight | ✕Can’t tell you what the data means for your business | ✓Plain-language insight and next steps |
| Keeping up with Google, ChatGPT and Claude changes | ✕Can’t keep up with algorithm changes or how to compete on each | ✓Adapts the strategy for each platform |
| Someone accountable for results | ✕No one | ✓A team that owns the plan |
| Cost | About $20 a month | Plans from $995 a month |
Key takeaways
- AI tools are excellent assistants. We use them for research, outlines, drafts, technical fixes and tracking on every account.
- They can only remix what already exists. AI engines reward the opposite: in Princeton-led research, adding citations, quotations from credible sources and statistics made content 30% to 40% more visible in AI answers.
- AI answers are built on other people’s coverage. University of Toronto researchers found AI search relies far more on earned media, such as independent publications, reviews and comparison sites, than Google does. That coverage has to be earned by people.
- Search never sits still. When Google shows an AI summary, people click a regular result about half as often (Pew Research Center). Someone has to keep adapting, every month.
- The question we hear on every sales call
- Where AI tools shine, and why we use them
- Content that doesn’t exist yet, not AI slop
- Credible mentions you have to earn
- Optimization that never stops
- When ChatGPT or Claude is enough
- Frequently asked questions
The question we hear on every sales call
“Why would I pay an SEO agency when I have ChatGPT and Claude?” We hear some version of that question on almost every sales call, and it’s a fair one. AI can write a blog post in thirty seconds. So what exactly does an agency do that a $20-a-month subscription can’t?
Our answer: a lot. And we say that as a team that uses AI every day. AI is very good at rearranging what has already been published. It can’t create new information, it can’t earn the trust of an editor or a podcast host, and it can’t own your results month after month. Those three things are where rankings, AI citations and customers come from.
Where AI tools shine, and why we use them
Let’s start with what ChatGPT and Claude do well, because they do a lot well. They cluster keywords, draft briefs and outlines, write serviceable first drafts, suggest title tags and meta descriptions, generate schema markup, help make technical fixes, track rankings and AI visibility, and explain any SEO concept in plain English. An agency that isn’t using these tools in 2026 is slower and more expensive than it needs to be.
Crowdcreate is human-led and AI-powered. AI gives us speed and structure across research, drafting and reporting. Our strategists, writers and editors supply what AI can’t: judgment, relationships and original insight. We even built Expert Press, our AI marketing platform, around that split. The rest of this article is about the human half.
1. Content that doesn’t exist yet, not AI slop
A language model writes by predicting the most likely next words based on what has already been published. That’s what makes it fluent, and it’s also its ceiling. Ask ChatGPT how a fintech startup should lower its customer acquisition cost and you’ll get a sensible list. So will everyone else who asks, and thousands of companies are publishing that same list right now. That’s AI slop: fluent, generic and interchangeable.
What an AI tool writes
“To lower CAC, focus on content marketing, optimize your funnel, invest in referrals and test your ad creative.”
Accurate, generic, and already on thousands of other blogs.
What your expert knows
“When we paused branded search for 60 days, signups held steady. Partner webinars made up the difference at a third of the cost.”
Specific, first-hand and quotable. Only your team can say it.
Illustrative example. Every company’s insight is different, which is the whole point.
Search engines and AI engines are looking for the right-hand column. Google says it rewards helpful content that shows first-hand experience and expertise, however it’s produced, and its spam policies call out scaled content abuse, meaning mass-produced pages made to manipulate rankings, with generative AI named as a common method. In Princeton-led research on generative engines, the best-performing changes were adding citations, quotations from credible sources and statistics, which lifted a page’s visibility in AI answers by 30% to 40%. Classic keyword stuffing did 10% worse than making no changes at all.
So our process starts with a calendar invite, not a prompt. We interview your founders, executives and most experienced operators, then turn what they know into articles, expert commentary, data pieces and quotes that go out on a regular schedule. AI helps us move faster on structure and editing. The substance comes from your people, which is exactly why competitors can’t copy it.
In practice, that means monthly interviews with your leadership team, articles built on first-hand insight, real numbers and quotes, timely pieces that react to news in your industry while it’s still news, and every fact checked and sourced before anything goes live.
2. Credible mentions you have to earn
The second thing an agency does is get other people talking about you: features in reputable publications, interviews on the podcasts your buyers listen to, and quotes in industry blogs and newsletters. That has always mattered for Google. In AI search, it’s becoming the main event.
Research from the University of Toronto shows why. The researchers compared the sources Google and AI search relied on for the same questions. AI search leaned heavily on independent publications, reviews and comparison sites, cited brands’ own pages far less, and almost never cited social posts. Their advice to companies was blunt: “dominate earned media to build AI-perceived authority.”
Here’s the part AI can’t do: earn that coverage. Somebody has to find the right journalist, editor or host, pitch a story they actually want, follow up, get your executive on a call, supply an expert opinion on deadline and then do it all again the next week. A chatbot can draft the pitch. It can’t build the relationship, take the call or be trusted by the person on the other end.
That’s what our digital PR team does every day. We can get clients featured in outlets like TechCrunch, Digital Trends, USA Today, Benzinga, Entrepreneur and IBTimes, plus the podcasts and industry blogs their customers already follow.
3. Optimization that never stops
SEO keeps getting better, and harder, because the platforms never stand still. Google ships core and spam updates, AI Overviews and AI Mode change what a results page looks like, and AI engines keep shifting which sources they trust. A strategy that worked last spring can quietly stop working this fall.
Fewer clicks when Google answers first
Share of Google search visits where people clicked through, based on the real browsing of 900 U.S. adults
Clicked a regular result, no AI summary on the page
Clicked a regular result, AI summary on the page
Clicked a link inside the AI summary
A randomized experiment by researchers at the Indian School of Business and Carnegie Mellon found the same thing from another angle: Google’s AI Overviews cut clicks to outside websites by 38% on the searches where they appeared. Ranking alone now earns less traffic than it used to. You need to be the source the answer is built from, and you need to keep adapting as the rules change.
So every month we fix technical issues, track rankings and AI visibility (which questions mention you, which don’t, and who gets cited instead), refresh pages that are slipping, publish new expert content and report back in plain language. AI tools help us collect that data. What they can’t do is tell you what it means for your business, or how to keep competing on Google, ChatGPT and Claude as each one changes. We don’t promise a “#1 ranking,” because nobody can honestly promise that. We commit to the work, and we measure it.
JumpOrange · E-commerce SEO
Technical fixes, a keyword and content plan and steady publishing gave JumpOrange its best year ever in organic search.
Read the case studySoku AI · AEO, SEO & creators
For a16z-backed Soku AI, we benchmarked its category across ChatGPT, Gemini, Perplexity and AI Overviews and mapped 11 prioritized growth opportunities.
Read the case studyWhen ChatGPT or Claude is enough
If you’re pre-revenue, have no marketing budget or just need a few blog drafts, use the AI tools and get started. We’d rather you spend $20 well than $2,000 badly. An agency makes sense once there’s revenue to grow, your market is competitive, you want to be cited in AI answers and not just indexed, and nobody on your team has the hours to interview experts, pitch journalists and keep improving the site every month. Crowdcreate plans start at $995 a month, our most popular plan is $2,995, and we work on a three-month minimum because the results compound. See what’s included on our SEO, AEO and GEO services page.
Frequently asked questions
Can ChatGPT or Claude do SEO on their own?
They can handle parts of it well: keyword ideas, outlines, drafts, meta tags, schema, technical fixes and rank tracking. They can’t create original expert insight, land press and podcast placements, turn the data into human insight, or keep up with how Google, ChatGPT and Claude keep changing and what it takes to compete on each. Those are the parts that decide whether you get found.
Does Google penalize AI-generated content?
Not for being AI-generated. Google says it rewards helpful, high-quality content however it’s produced. What it acts against is scaled content abuse: mass-producing pages to manipulate rankings without adding value. Generic AI content rarely breaks rules; it just rarely stands out.
How do I get my company mentioned in ChatGPT and Google’s AI Overviews?
Earn coverage on the independent sites AI engines trust, such as publications, review and comparison sites and respected industry blogs, and publish original content with real quotes, data and sources. University of Toronto researchers found AI search relies far more heavily on this kind of earned media than Google does.
How long does SEO take to work?
New expert content and press coverage can start showing up in AI answers fairly quickly, because AI engines favor fresh, credible sources. Meaningful ranking and traffic growth usually takes three to six months, which is why we work on a three-month minimum and report on progress every month.
Want content AI can’t write and coverage AI can’t earn?
Tell us about your business and we’ll come back with a plan: what to publish, where to get featured and what to fix first.
Book a strategy callSEO, AEO & GEODigital PRAbout this article
Written by the Crowdcreate SEO and Digital PR team. Crowdcreate has been a growth marketing agency and advisory since 2014, helping brands including Anker, Lenovo, TP-Link and IBM, and startups backed by investors such as a16z and Goldman Sachs, get found on Google and in AI search, earn press and creator coverage, and reach customers and investors.
Update log
November 2026: Rewritten with new research from Princeton and the University of Toronto and new Crowdcreate case studies.
November 2026: First published.
Sources
- Google Search Central, “Google Search’s guidance about AI-generated content”
- Google Search Central, “Spam policies for Google web search”
- Aggarwal et al., Princeton University and IIT Delhi, “GEO: Generative Engine Optimization” (KDD 2024)
- Chen, Wang, Chen and Koudas, University of Toronto, “Generative Engine Optimization: How to Dominate AI Search” (2025)
- Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results” (July 2025)
- Search Engine Journal on Agarwal and Sen’s randomized field experiment (April 2026)