It knows where you live 🎃

Hello marketers. Welcome to AI Marketing School, where we dish out the latest and greatest in AI-powered marketing.

In this week’s ChatGPT Search mega issue:

1. AI Marketing Update: OpenAI drops ChatGPT Search

2. The Stack: Using ChatGPT Search

3. Consultant’s Corner: AI saves $1 million on a marketing campaign

Onwards!

AI MARKETING UPDATE

OpenAI’s ChatGPT Search shakes the SERPs

Perfect timing for the newsletter this week with OpenAI dropping its ChatGPT Search feature on October 31st.

AI-driven search tools like Google SGE and Perplexity AI offer new methods for searching the web, providing complete narrative answers rather than a list of links.

ChatGPT is taking this same route, and it’ll prove a very useful intriguing addition to the company’s growing portfolio — with both challenges and opportunities for marketers.

Here’s an overview of how ChatGPT Search works and what it brings to the table:

Search flexibility

First off, ChatGPT Search is in a different league to ChatGPT’s old ‘browse with Bing’ feature, which will likely soon be retired.

ChatGPT now automatically decides when a question requires web data, but users can also trigger searches manually with a globe icon. If you search for something topical or time-based, it’ll usually kick in.

Real-time, source-linked answers

ChatGPT provides answers with direct links to sources from what looks like a) a curated list of sources and b) supplementary information from Bing’s index.

OpenAI has been busy partnering with all sorts of media companies such as AP, Reuters, Le Monde, The Financial Times, and Condé Nast, so its repertoire is quite broad, but it will still probably lean on certain voices and opinions.

Many other sites besides current affairs and media seem to appear in ChatGPT Search, from review sites to charities and government websites — all sorts.

However, right now, it’s difficult to tell exactly what ChatGPT’s inclusion criteria are and where exactly it draws its data.

Impact on search engines

As we discussed in March, tools like Google SGE and Perplexity set the stage for an answer-driven search experience, designed to respond directly to user questions instead of directing users through a series of links.

In my tests, ChatGPT Search is amazing at tailoring responses to different searches, including recommending products, photos, graphs, and video embeds. It’s much more diverse and comprehensive than anticipated.

This hints at broad SEO, traffic, and marketing impacts, too.

Chrome plug-in

As part of the roll-out, OpenAI has released a Chrome plugin that turns your default search engine to ChatGPT.

When you type a query in your URL bar, it’ll open ChatGPT and search it there instead.

Be ready to hear a lot of “Google Killer” talk about the new ChatGPT.

How to optimize for AI Search

It’s impossible to tell right now whether ChatGPT Search draws info only from a curated list of sources, and whether SEO (if you can call it that) factors into it much at all.

Based on lessons we’ve seen from SGE and Perplexity (and explored in this past edition), AI prefers clear, quick answers, so keep your intros short and get straight to the point. Answer first, detail later.

ChatGPT’s answer-first approach also values hard facts, so integrate sources and concrete data where possible. Citing reliable sources is best practice all around.

The honest truth? No one is going to know exactly how this behaves, or whether smaller sites outside of its curated sources have a chance of featuring in ChatGPT’s search outputs.

And it knows where you live

Something for spooky season — ChatGPT Search knows where you live and sometimes tries to give (very) location-specific answers.

I found ChatGPT Search often mentions your town/city based on your public IP address (or so it says). However, my small town is not deducible from a public IP.

When quizzed about how it knew where I was using the app from, ChatGPT replied: “I don't actually have access to your IP address or any precise location data. My previous mention of [Location] was simply a mistake, and I don’t have access to any personal information unless you directly share it.”

There is virtually no chance I shared my location with the specificity required for it to give me localized results at the snap of the fingers.

Maybe this won’t surprise some people, but it sure surprised me.

THE STACK

Strategies for using ChatGPT Search

ChatGPT Search integrates research directly into content creation processes, addressing issues with data accuracy and hallucinations in the process.

This means less back-and-forth between tools and more ‘in-the-box’ creativity.

Here are some marketing-related strategies to explore:

Creating news and other topical, up-to-date content

Below is an article created based on the crypto query above. It took about 1 minute to create from start to finish.

Compare that to collating data from multiple sources manually and writing them into a coherent piece. It’s far more detailed and up-to-date than ChatGPT’s previous Browse With Bing search function.

You could even create the feature image with DALL-E and have the post uploaded within minutes.

Social media uses

ChatGPT Search is excellent at rounding up information, e.g., ‘give me the latest [industry] insights from this month.' This offers a starting point for creating blogs, social media posts, and even video scripts.

For the below query, it even returned a YouTube video embedded into the chat, which was a little surprising.

Local marketing

As noted earlier, ChatGPT Search is spookily good at localizing searches. This is excellent for anyone who needs to research local information for marketing and content purposes.

Again, integrating the research into the creation process means you can spin up posts, blogs, social media info, etc, all from one single interface.

All sorts of competitor and demographic analysis

ChatGPT Search is also great at competitor analysis, demographic analysis, and other forms of marketing research that require salient, up-to-date information.

Some insights are obvious, and you’d find them from normal searches; others are less obvious and would take a fair amount of work to unearth.

There will no doubt be other uses emerging in the weeks ahead.

One limitation: No access to social media is ChatGPT Search’s biggest blind spot. If Microsoft and OpenAI could sort this out, it would add a fifth dimension to this tool — though it’s difficult to envisage it happening.

Being able to explore trends and topics as they emerge on social media pace would be insane. Only xAI’s Grok can do this right now, but it lacks the teeth and features of ChatGPT and is confined to X (no broader internet access).

CONSULTANT’S CORNER

An AI marketing campaign that worked and saved $1 mil

We’re back with another AI campaign story, and this time it’s not a horror show.

IT management platform Atera just launched a fully AI-generated video campaign, saving up to $1 million in production costs.

Given the mixed results of AI-driven ads we’ve seen lately, this B2B brand played it smart by balancing automation with creative direction.

The concept? Visualizing the dreams of IT managers. Picture a work-from-the-beach fantasy, UFOs blowing up stubborn printers, and even astronauts on laptops. It’s so smartly absurd that it works with the AI visuals.

With tools like Sora, Midjourney, and Runway, Atera’s creative team leaned into AI for these visuals, trimming what would usually be a months-long production down to just four weeks.

This wasn’t a totally hands-off process, though. As Elad Gaizler, Atera’s creative marketing manager, explained, “The human element started with the concept development and scriptwriting… From that point, we allowed AI to interpret and transform the concept into visual form.”

So why did it work?

1. Audience Matters: Atera’s B2B audience — tech-savvy and likely arguably more tolerant of AI than typical B2C markets — meant this AI-driven campaign was less of a risk. For IT managers, a high-concept ad produced quickly and efficiently resonates; the novelty factor is a bonus rather than a distraction.

2. Human Guidance + AI Execution: Atera didn’t just hand everything to AI. “We remained actively involved throughout the process, guiding the results to ensure they aligned with our vision,” Gaizler noted. This balanced approach allowed AI to generate wild visuals while the team stayed in control, keeping the tone in line with Atera’s brand.

3. Huge Savings, Smart Use of Budget: The real selling point here is the cost. By leveraging AI, Atera achieved a high-production look for a fraction of the traditional price. And saving $1 million while keeping the creative fresh and edgy? A great result.

So, if you’re eyeing AI for your next campaign, take a note from Atera: it can work, especially in that B2B context where you’re less likely to face the public’s wrath.

And sometimes, the best results come from letting AI go bold, as long as you’re there to steer the ship.

Hope you enjoyed this week’s issue. If you missed it last week, you can read it here.

If you found it useful, please recommend it to a friend or colleague.

Until next time. Happy marketing.

—The AI Marketer