Hello Reader
People often ask me which AI platform they should use. My answer is usually the same: start with the one you have access to.
Recently, I ran an informal poll asking marketers whether they use ChatGPT or Claude. The vast majority of those who responded use both.
That surprised me.
Getting an AI platform familiar with your work takes effort. You explain your business, upload reference material, build projects and teach it what good work looks like. Using two platforms means you have more of that context to maintain.
But marketers are finding reasons to do it.
While ChatGPT remains the largest standalone AI website by monthly visits, the market has become significantly more competitive. According to Similarweb, its share of visits has steadily declined as rivals such as Gemini and Claude gain momentum.
Similarweb also shows there’s a lot of overlap between the tools — 29 million people are using all three.
I was curious about this usage so I asked a number of marketers in my network why they’re using both and what for.
Their answers included clear preferences, extra capacity and the occasional need to walk away before losing their temper.
Why Marketers Keep Both
For fractional CMO Chris Arden, the division is straightforward.
ChatGPT handles business development, including outreach, positioning and thinking through his go-to-market approach. Claude handles coding and building systems.
“ChatGPT is better at sounding like me on business development writing. Claude is better when I’m at wiring systems and code.”
Jenn Steele, Cofounder of SoundGTM, prefers ChatGPT’s writing, even after giving Claude comparable guidance.
“ChatGPT has adopted my writing patterns better, even though I’ve given Claude just as much guidance and examples,” she says. “Plus ChatGPT’s cynical mode speaks to my cold, bitter GenX heart.”
She shared a recent example of asking both platforms to write her speaker bio for an upcoming conference.
“Claude was a disaster,” she says. “But ChatGPT did really well, especially when I pointed it at the site that had the bios I needed to mimic.”
Tiffany Teixidor, VP of Marketing at Educate 360 has a similar preference for ChatGPT’s language work, with a different role for Claude.
“ChatGPT is my strategy and language partner,” she says. “I use it for brand voice, positioning, audience psychology, message development and pressure-testing ideas.”
She turns to Claude to translate her thinking into visual concepts, layouts, presentations and other creative outputs.
Fractional CMO, Lauren FitzHugh primarily uses Claude, with specific exceptions.
“I tend to use ChatGPT for quick image exploration and feedback on audio files, including tone and cadence.”
For many of the marketers I talked to, maintaining two systems was all about using the right platform for the right task, but a few also use it to extend their work.
Justin Parnell is a FDE (Forward Deployed Engineer) working with marketers and his main reason for using both is credit capacity. “When I exhaust my Claude subscription, I go to OpenAI,” he says.
Jenn has another reason. “When I’m about to lose my temper with one of them, knowing I can switch to the other helps me keep my calm.”
Anyone who has spent twenty minutes asking an AI tool to stop making the same mistake will likely feel the same.
Give New Models A Test
Having both platforms available also gives you a way to try new capabilities.
Online discussions are divided about switching whenever a new model arrives.
Some users want access to the strongest option for each task; others prefer to keep working with a model that already meets their needs.
Charlie Treadwell, CMO at Elisity sees value in keeping up with both. “If you have the time, I think it’s important to be familiar with the capabilities of both models.”
You can make the comparisons easier by preparing a test project now.
Set it up with the context and instructions the work requires, including a brief, source material, brand guidance and an example of an acceptable result.
Choose two or three recurring tasks you can judge well. You might use an article edit, a campaign analysis and a messaging brief built from customer interviews.
Run those tasks with your current model. Save the prompts, outputs, model name and date, along with a short note about what you had to fix.
When a new model becomes available, run the same tests in fresh conversations within the prepared project. Keep the instructions and source material unchanged. For a comparison across platforms, prepare matching projects in each.
Look at the outputs side by side.
- Did the new model follow the brief more closely?
- Did it catch something useful?
- Did it reduce the editing?
Once you have the answers to those questions you’ll know whether or not it’s time to switch.
Let Them Challenge Each Other
While it’s most common for marketers to assign each tool distinctive tasks, there are a handful who are pitting Claude and ChatGPT against each other.
“I frequently toss deliverables from one to the other for a second opinion,” says Tiffany. “Just like I would with my human team members.”
Because both have access to the same reference material, she finds the feedback relevant while the perspectives differ.
Charlie takes the handoff further.
In a recent example he asked Claude to search SharePoint and prepare design documents, then handed those to ChatGPT (Codex specifically) to update the briefs in Adobe Illustrator and InDesign.
“ChatGPT navigated to our product page, took screenshots, cropped them and used Illustrator to turn them into nicely designed product shots,” he says. “It also updated the copy and checked for text overruns.”
Charlie also uses one model to review the design while another continues improving it.
Which for Jenn, doesn’t always go according to plan.
“Sometimes, if I try to have one build on the other’s work, it becomes a giant, smoking pile of crap,” she says. “When that happens I usually have to start from scratch.”
Don’t Skip the Setup
Chris keeps his two tools focused on separate responsibilities.
“I don’t try to make them identical. I care more that each one does its job well than that they match.”
Once he establishes that split, he finds the overhead fairly light, but he recommends only adding a second platform when you have a task that your current tool handles poorly.
If you do decide to use both, make sure you dedicate enough time to complete the setup.
Jenn’s experience shows the cost of leaving that unfinished.
“There are some projects I haven’t fully fleshed out, and sometimes that bites me when I’m trying to get something done.”
Every AI platform needs the same core inputs to deliver useful results. Make sure you have these ready:
1. A clear task & outcome
Explain what you need to produce, who will use it and what a good result should accomplish.
“Write a bio” leaves a lot open. A bio for a conference speaker page has a specific audience, purpose and format.
2. Useful reference material
Supply the facts, examples and source documents ChatGPT or Claude should draw on.
Jenn’s example biographies gave ChatGPT something concrete to follow.
3. The context that supports the output
Include the decisions already made, the language you use, the constraints and the information the tool should take into account.
Tiffany gives both platforms a substantial body of reference material. That includes brand guidelines, positioning, audience profiles, product information, campaign history and key decisions.
“The most important thing I’ve learned is to build a strong knowledge base,” Tiffany says. “Take time to create it and keep it updated.”
Want To Up Level?
Keeping reference material current across several tools gets more involved when you’re using them across your team.
Justin advises his clients to store their context outside the individual AI platforms using Notion, Supabase and Github.
In his setup he has what he calls a “layered context stack” where each layer directs a different part of the system, has a different owner, and refreshes on a different rhythm.
Here’s what that looks like:
Charlie uses a plugin marketplace stored in GitHub to manage skills, agents and connections to other systems. GitHub stores the files and tracks changes and the marketplace packages those resources so the people on his team can install what they need.
“Building that foundation prepares you to switch platforms when you need to,” he says.
Want me to break down the system architecture and setup behind this?
Hit reply and tell me—I'll cover it in a future edition.
Getting More from Your AI
Marketers will get more from AI when they give each model the right job and enough context to do it well.
For me, that currently means using ChatGPT for rapid brainstorming, messaging and business development writing, and Claude for complex synthesis, long-form content and system wiring.
Keep your brand context, project instructions and reference materials in one central repository. You can use the same source material across both platforms and carry it into any model you use next.