Blog

AI tools like ChatGPT, Claude, and Letaido have gotten incredibly good over the past year. Tasks that used to require a software engineer can now be done by anyone with a bit of curiosity.

Here are 37 practical ways marketers are actually using AI right now. Every single one comes from someone who’s done it, documented the process, and gotten real results. No theory—just stuff that works.

---

SEO (Search Engine Optimization)

SEO is full of tedious, fiddly work—spreadsheets, data exports, complicated analysis. That combination of repetitive and complex is exactly where AI shines.

A quick warning: Anyone promising to automate your SEO for $9 a month is selling snake oil. AI isn’t a substitute for knowing what you’re doing. It’s a tool that still needs a skilled person driving it.

1. Get a Google-quality rating in minutes

Dan Hinckley built an AI agent that reviews your page the way a Google quality rater would. You give it a keyword and a URL. It reads Google’s 168-page rater handbook, turns it into a checklist, opens a real browser, searches for your keyword, clicks through your page, scrolls through it, checks your About page and footer, and writes up a detailed report. It even records a video of everything it did.

The handbook is the closest thing we have to Google telling us what a good page looks like. But most of us aren’t going to read 168 pages. This gives you the verdict in minutes.

2. Do an afternoon of keyword research in 20 minutes

Our VP Marketing built a keyword research tool because he was tired of sorting keywords by hand. Give it a niche like “coffee,” and it expands into seed keywords, pulls real search volume data, reads the search results to judge intent and difficulty, and categorizes everything into “Go,” “Maybe,” or “Skip.” It even tells you if a search result page is full of videos or Reddit threads before you write a single word.

In one run, it graded 8,437 keywords into 17 clusters and marked 1,809 as “Go.”

3. Turn site audit issues into developer-ready tasks

We connected our site audits straight to GitHub. Every Sunday, it pulls the latest audit, ranks issues by severity, and opens a pull request with a checklist of high-priority problems—indexability issues, broken pages, broken links.

Developers don’t have to figure out what’s wrong or which URLs are affected. They just review and approve the changes. Easy.

4. Add FAQs to old posts, get 32% more traffic

Steven Macdonald ran a real experiment. He took 21 old posts, used ChatGPT to find questions readers would still have after finishing each article, added FAQ sections with answers, and published.

Traffic rose 32% on average—about 16,000 extra visits and 3,500 new keyword rankings. His control group (posts he didn’t update) dropped 4% over the same period. That’s the difference between “I think this worked” and “I know this worked.”

5. Build an internal linking engine (even if you know nothing about coding)

I built a tool that suggests internal links for any blog post—and I started by asking ChatGPT’s Deep Research to teach me the math behind it. I had never worked with any of this before.

Here’s the simple version: it reads every article on the blog, turns each one into a string of numbers that represents what it’s about, then compares any URL against all the others and finds the 20 closest matches. It finds pages about the same subject rather than pages that share keywords—and those are usually very different lists.

Internal linking matters because it helps Google find your pages, passes authority around your site, and tells Google what each page is about. Google’s John Mueller called it “one of the biggest things you can do on a website.”

The hard part is holding hundreds of articles in your head at once. This tool does that for you.

6. Let AI write the internal link into your page

John Iwuozor wrote a Python script that doesn’t just find a good internal link—it actually writes it into your article. It reads your other pages, finds the closest match, scans your article for a phrase that would make good anchor text, and drops the link in.

Finding the right page is half the job. The other half is opening the article, reading through it, choosing the right sentence, and editing. That second half is why internal linking backlogs never shrink.

7. Turn Search Console data into a prioritized action plan

Rudiger Dalchow built a tool that reads your Search Console exports and tells you exactly what to write next. It groups your pages into topics, figures out what people want from each one, identifies pages that are competing with each other, and writes briefs you can pass straight to a writer.

Instead of just telling you what happened, it tells you what to do.

---

AI Search Optimization

You want your business to show up when someone asks ChatGPT about what you sell. Figuring out how is harder than it should be—there’s lots of jargon and snake oil.

But plenty of what actually works is straightforward. Here’s what people are doing:

8. See your page the way an AI sees it

Aimee Jurenka built a tool that takes your URL, captures the raw HTML and a screenshot, then asks GPT how the page reads to a language model. It gives you a score and feedback on whether your content is clear enough to be extracted, worth citing, and whether anything important is hidden behind JavaScript.

That last one catches something people miss—a page that looks fine in a browser can be almost empty to a crawler that doesn’t execute your scripts.

9. Map “fan-out” queries against your content

Chris Long wrote a script that checks whether your site answers the questions an AI assistant is actually asking. When you ask an assistant something, it rarely searches for exactly what you typed. It breaks your question into several narrower ones.

The script pulls out those narrower questions, scores each one against your pages, and tells you where your content falls short. A page that covers four of six questions behind a topic gets used less than one that covers all six.

10. Rewrite pages based on what people actually asked

Wil Reynolds uses Bing’s free webmaster tools to see which of his pages are being used in AI answers—and what people asked to get there. He collects the real questions that led AI assistants to his page, then rewrites his page to actually answer those questions.

Instead of guessing, you’re working from the actual queries that brought people there.

11. Build page templates from what AI actually quotes

Adina Timar tracks 460 prompts across ChatGPT, Perplexity, Gemini, and Google AI Mode. She studies which pages get quoted, on which platform, for which prompt, and how each assistant structures its answers. Then she builds her page templates around what the data actually shows.

Most people guess. She replaced guesswork with a record of what got quoted. Her key insight: different prompts favor different page types, and the winning pages are the ones with honest verdicts—where you admit where a competitor beats you.

12. Decide what to write next by tracking AI citations

Ross Hudgens’s agency starts with the prompts that actually lead to sales—”what’s the best X” questions covering everything they sell. They track which pages get cited most often across AI answers. When they see competitors being cited, they know that spot is winnable.

Then they check those same topics against Google demand. Anything cited heavily by AI and searched heavily on Google goes to the top of their priority list.

---

Content Marketing

Ask AI to write a blog post and you get slop. No personality, no original thought, no research.

But treat content marketing as a multi-step process (which is how skilled writers already work), and AI becomes remarkably useful for many of those steps: research, briefing, outlining, editing, fact-checking, internal linking, formatting.

13. Get a publishable draft in about two hours

I broke our editorial process into small steps—topic selection, briefing, outlining, structural editing, drafting, line editing, internal links, metadata—and wrote each step down. Those documents live in a ChatGPT project.

The trick is the number of steps. One prompt gets you slop. Ten narrow prompts, each doing one job with someone checking the output, gets you something publishable. You still need a competent marketer driving it.

14. Learn to build web pages yourself (even without coding experience)

Kelsey Libert spent a month learning Cursor—a code editor with AI built in. She described what she wanted in plain language, and it wrote the code. In a month, she built 30 landing pages for her company’s agent products.

Landing pages are the classic marketing bottleneck. You know exactly what the page should say, then wait weeks for someone else to build it. Being able to ship it yourself removes a handoff that costs most teams more time than the writing does.

15. Run a real data study without hiring an engineer

Mateusz wanted to know if a company’s organic traffic tracks its share price. He had AI write the code to pull three years of traffic and stock data for every ticker on the Nasdaq, run correlations, and draw charts. Pretty much all the number crunching in that article was AI.

The study got better when a real data scientist took over the analysis—but AI got him from no data to a working dataset. That’s the part that usually kills a research project before it starts.

16. Stop ChatGPT from writing like ChatGPT

Myriam Jessier uses a custom instruction she calls “Absolute Mode.” It tells the model to drop emojis, filler, hype, conversational warm-ups, and unnecessary summaries. It stops asking follow-up questions and ends the reply the moment it’s finished.

Most editing AI copy is just deleting things. Fixing that at the instruction level rather than sentence by sentence is the best ten seconds you can spend.

17. Talk through an idea and get a deck back

Wil Reynolds records himself explaining a concept, transcribes it, feeds it to an LLM to draw out a story arc, then uses NotebookLM and Gamma to turn it into deck variations.

What comes back isn’t a finished deck—it’s a rough shape he edits. But talking is much faster than writing, and the slide-making parts that eat an afternoon are already done.

18. Make a 1,000-article refresh backlog manageable

We have about a thousand articles on this blog. Keeping old ones current is some of the highest-return work we do. A post published three years ago can still be the best thing on the internet about its subject and still be losing traffic because a statistic is outdated.

It matters for AI search too—AI assistants cite noticeably fresher pages than Google does. Our Update Pipeline does the checking. Give it a URL, it looks for stale statistics, notes where we describe our product as it used to work, and compares the article against what currently ranks. Then it shows you old and new side by side, and you accept or reject each change. Nothing goes live unread.

19. Find statistics your competitors haven’t discovered

Steve Toth built a Claude project that goes hunting through PDFs, Word documents, and PowerPoint decks rather than ordinary web pages. About ten minutes later, it returns roughly 30 statistics with sources and file names.

The statistics everyone quotes are sitting on web pages—easy to find. The ones nobody quotes are sitting in the appendix of a PDF nobody opened. Those original numbers are one of the few things left that still earn links.

20. Query a pile of expert quotes instead of sorting through them

Getting expert quotes makes an article worth reading. But you post a request, and a few days later there are dozens of replies in your inbox—most unusable, and you have to open every one.

Mateusz uses NotebookLM to handle the pile. He saves email notifications as PDFs, uploads the lot, and waits a minute while it reads them. Then he can ask questions of the whole set at once: “Show me the most unconventional tips from business owners, and tell me which file each one came from.”

It can hold far more text than most chat tools, and it only answers from the documents you gave it—with references back to the source.

21. Turn every webinar into three pieces of content

Ryan McCready chained four Zapier automations together. The first takes a webinar recording and produces three blog outlines, a guide to the best clips, and social briefs. The second grows approved outlines into 1,200-word drafts with sources and quotes. The third listens to sales calls and files every pain point into a database. The fourth does the same with customer calls.

Every company records these calls. Almost nobody listens to them again. The database of complaints is what makes this more than a time-saver—most content teams have no record of what customers actually grumble about.

---

Marketing Analytics

Reporting is the most automatable job in marketing. It’s also nobody’s favorite part of the month—a clear sign a machine should be doing it.

AI can pull exports, join them up, and build tables. What it still can’t do is tell you what it all means. That part stays with you.

22. Turn a full day of monthly reporting into a scheduled job

Our monthly blog performance report now runs automatically on the 2nd of each month. It pulls from Google Search Console and Ahrefs Web Analytics, creates KPI tiles, a 12-month trend chart, subfolder splits, winners-and-losers tables, daily anomaly callouts, and paginated post lists.

It also drafts 6-10 candidate analysis bullets for whoever writes the commentary, so they’re editing rather than starting from a blank page. What used to take most of a day now runs automatically.

23. Check whether your content is actually on-topic

We ran a semantic audit of this blog to find out how many articles cover our core topics—SEO, link building, content marketing—and how many cover topics we’re not an authority on.

The process: turn every article into a string of numbers representing what it’s about, average them all together to get a mathematical answer to “what is this site about,” then measure how far each article sits from that center.

The pattern was clear: core pages get about twice the organic traffic of the far ones. This puts a number on what that drift costs you.

24. Pipe crawl, analytics, and Search Console data into one audit

Chris Long documented a content audit built by Ian Lurie that reads your crawl, analytics, and Search Console numbers together. Screaming Frog, Google Analytics, and Search Console all feed into Claude through Zapier. Claude reads all three together and looks for pages losing traffic, pages answering the wrong question, and pages that no longer fit the site.

Many content problems are invisible in any single report. A page can look healthy in analytics while quietly losing impressions. You only catch that when the three datasets sit side by side.

---

Social Media and Community

Social media is hard for two reasons: people mention your brand in more places than you can keep up with, and there’s always more you could be posting than you have time to write.

AI is good at sorting and drafting—not at publishing. Getting a reply wrong on social is more expensive than getting a report wrong. So let AI find mentions worth answering and give you a first draft, then have someone who knows the room decide what actually goes out.

25. Find the Reddit conversations worth joining

The lemlist team spent three months testing Reddit. They started 12 threads, joined 28 existing discussions, and replied to 13 places where someone mentioned their brand. Traffic from Reddit went up 22%, and sign-ups were 20% higher than average.

The part AI handled was the looking. Reddit is enormous, and valuable conversations are buried. They used AI to surface the handful of relevant threads and give them a starting point. Everything after that was a real person.

26. Learn how your customers actually describe their problem

Britney Muller walks through how to pull audience insight from community platforms. Her point: the conversations you most want to read are the ones nobody has time to find. That’s now a solvable problem.

The version she recommends uses a platform’s free API to collect discussions mentioning your brand or subject, then has a model read through and pull out recurring complaints, questions that come up again and again, comparisons between you and competitors, and the exact words people use.

What comes back is worth more than a sentiment score. It’s the language your customers use when they don’t know you’re listening—which is usually nothing like the language on your website.

27. Get five social drafts in your own voice for every article

I built a social post generator because writing the LinkedIn post was the bit I put off longest. It checks the blog’s sitemap for anything new under my byline, reads the article, and writes five different drafts.

One leads with the most surprising thing, one opens with a story, one asks the question the article answers, one lists takeaways, and one argues against common belief. I don’t know which angle I want until I see a few—picking between options is much easier than starting from nothing.

The part that made it usable is that it writes in my voice. I marked up my best-performing posts as style examples. Once I’ve picked a draft and edited it, it pushes to my scheduling queue.

---

PR and Outreach

Outreach lives or dies on timing and relevance. Being the third person to pitch a journalist is worth nothing. The job involves watching thousands of stories a day across feeds nobody has time to read, then writing to each contact as though you’d read their work.

AI-generated outreach at scale is already why many journalists have stopped reading their inboxes. So these processes stop at the same place: they do the watching and the first draft, and leave you to decide what actually gets sent.

28. Monitor thousands of news stories a day for PR opportunities

Mark Williams-Cook wrote a Python script running a local AI model across thousands of daily news stories from RSS feeds. It matches each story against client interest profiles and sends qualified leads to Slack for the PR team.

Running the model locally makes the volume viable—no per-call API bill when you’re screening thousands of stories a day.

29. Write follow-up emails from call transcripts in five minutes

Andy Chadwick feeds sales call transcripts into a ChatGPT workflow trained on his tone, product details, and common objections. It drafts a follow-up reflecting that prospect’s specific problems. He edits and sends.

Composition went from 45 minutes to 5. He reports a 90% close rate on the high-ticket follow-ups he uses it for.

30. Run a cold outreach campaign that sounds like you wrote it

Sam Oh paid an agency to do outreach. They sent 100 emails and got three replies—two asking to be removed. So he built his own.

It runs as a team of agents with a manager. One agent finds prospects and pulls data on each, a second checks if their traffic growth is real, and a third writes emails and leaves them in his Gmail drafts. He runs it all from Slack.

He set it going as he left for the airport. By the time he reached his gate, there were 74 drafts. Of 54 sent, 11 came back and 7 turned into booked meetings.

31. Turn a data study into a press release automatically

We built a Press Release Generator that takes a blog post URL or product feature note and returns a formatted press release. The plan is to point it at our data studies so every new study comes with a release already drafted.

Original research is how we get picked up by the press—The Atlantic, CNN, BBC—and that only happens if someone pitches it promptly. The writing wasn’t the hard part, but it was where things stalled. Having a draft ready makes promotion happen consistently.

---

Product Marketing

Launching a product means writing a landing page, an email, a video script, a sales deck, and a one-pager—and they all have to say the same thing. They usually don’t. Different people write them in different weeks.

32. Generate an entire launch package from one product brief

Andrei built a GTM Generator that turns one rough brief into six things: a tidied-up brief, a landing page, a video script, a promotional email, a near-print-ready flyer, and a final check.

That last step reads all five documents side by side and tells you everywhere they disagree—down to the landing page promising “10x faster” when nobody claimed that in the brief.

33. Reverse-engineer a competitor’s paid campaigns into your own ads

Our Paid Ads Campaign Builder takes a competitor’s domain and gives you a set of ads to run against them. It pulls the keywords they pay for and the pages they send traffic to, reads those pages to see their actual promises, groups keywords into themes, points out ones you’re not bidding on, and writes Google Search ads within character limits.

Paid search is the one place a competitor can’t hide what they believe is worth money. Every keyword on that list is one they’ve tested and kept paying for.

34. Set up an entire webinar from a title and a date

Constance built a webinar automation app. She types in a title, a date, and who’s speaking, and it sets up everything: Zoom event, landing pages, promotional copy, paid promotion booking, slides, script, and post-event reporting.

It hands over tasks in stages so she only sees what she can do today. Running a webinar isn’t difficult—it’s just twenty small things that have to happen in the right order.

---

International Marketing

International marketing multiplies everything. Every article becomes seven articles. Most of the added work is mechanical rather than skilled—which is why it’s a good fit for automation and why it usually doesn’t get done.

The catch: translation and localization aren’t the same thing. A model will give you fluent Spanish that reads as though it were written for the wrong country, and you won’t notice if you don’t speak it.

35. Translate a blog post without breaking the links inside it

Erik and Taka built a translation pipeline that handles seven languages, each with its own tone guide, glossary, and translation rules. It also fixes the links—translate an article and the links inside it still point at English versions. This checks each link and swaps in the local version if it exists.

Before any of that, it does keyword research in the country you’re translating for, so the article goes after words people there actually search for.

36. Localize the charts and diagrams inside your posts

A translated article with English screenshots is only half translated. The international team built a translator just for visuals: upload a PNG, JPG, or PDF, pick a language and region, get the localized image back.

One model looks at the image and writes down what needs changing. A second one redraws it. You pick a region as well as a language, so Spanish for Mexico comes back different from Spanish for Spain, and prices, example web addresses, and names change to suit.

37. Run live subtitles at a conference for the price of a laptop

Professional simultaneous interpreters cost upwards of $2,000 a day. Taka runs a Live Interpreter on a laptop next to the stage instead, generating English and Japanese subtitles from the microphone feed.

Live speech arrives in half-finished chunks, so it waits and only puts complete sentences on screen. It reads from a glossary of product names first, so “AI Overviews” always comes out as the agreed Japanese term.

---

Final Thoughts

Plenty of people will tell you they’ve fired their marketing team and handed everything to AI. That’s nonsense.

AI is a tool, and it’s only as good as the person holding it. Every process on this list was designed by someone who knows marketing inside out. They know the workflows, they have taste and judgment, and they could get good results without AI—which is precisely why they get better results with it.

Stay skeptical. AI isn’t a magic box. But in hands like yours, it will make your work faster, some of it better, and some of it genuinely more fun.