Still using ChatGPT? Here's why Claude is worth the switch
Claude has quietly pulled ahead of ChatGPT. Memory, Projects, Google integration, and writing that doesn't sound like a robot — here's what you're missing.
Most people started with ChatGPT. It was there first, it worked well enough, and once something becomes a habit it tends to stick around.
But the gap between ChatGPT and Claude has been growing — quietly, without much fanfare — and if you haven't looked at Claude recently, you're probably leaving a lot on the table. Not because it's shinier, but because it actually works differently in a few ways that matter day to day.
Here's what's changed and why it might be worth making the switch.
It writes like a person
This is the most immediately obvious difference, and it's hard to overstate how much it matters if you use AI to help with writing.
ChatGPT has a voice. You know the one — slightly too enthusiastic, fond of bullet points, never quite sure whether it's a consultant or a motivational speaker. Ask it to write something and it'll often come back with "Absolutely! Here's a great option for you:" followed by something that sounds like a blog post written by a blog post.
Claude doesn't do that. It writes like it's actually thought about what you asked. Give it a bit of context about your tone and it'll match it well enough that you're editing rather than rewriting. For emails, blog posts, proposals, social posts — anything where the words need to actually sound like you — that difference is significant.
The sycophancy thing is real too. ChatGPT tends to agree with you. Claude will push back when you're wrong, which is considerably more useful.
It remembers who you are
This is the feature that changes everyday use the most.
Claude now has persistent memory — rolled out to all users in early 2026, including the free tier. It builds up a picture of you across conversations and carries it forward. Not just your name, but genuinely useful context: what you're working on, how you like to communicate, your preferences, your ongoing projects.
After a few weeks of regular use, you don't have to re-explain yourself every time you open a new chat. If you're working on something specific, it already knows the background. If you're asking it to write something, it already knows your preferred style. It's not magic, but it's a lot closer to an assistant who actually knows you, rather than one that meets you fresh every single morning.
You can go to Settings → Memory and see exactly what it's retained. Edit it, add to it, delete anything you don't want kept. Full visibility and control throughout.
Projects: a proper workspace
If memory is Claude learning about you over time, Projects is you giving it a focused brief for something specific.
Each project is its own self-contained workspace — its own chat history, its own knowledge base, its own instructions. You upload the documents that matter for that project, write a few lines about what you want Claude to keep in mind, and then every conversation you have inside that project starts from an informed position. You're not re-uploading the same files every time or explaining the background from scratch. It just knows.
On paid plans, projects handle large amounts of content without breaking a sweat — Claude automatically uses a technique called RAG (Retrieval Augmented Generation) when the knowledge base gets large, which lets it pull in the right context from your documents without getting overwhelmed. In practice this means you can load a project up with a lot of material and it'll still give you sharp, relevant answers.
You can set standing instructions at the project level too. Things like: always respond in a certain format, keep this constraint in mind, here's the background you need. It removes a lot of the friction from getting useful output.
I've got a project set up for each of my side projects — Pinx and Feedframer each have their own, loaded with an executive summary, marketing plan, branding docs, and any other reference material that's relevant. When I want to work on something — a landing page, a campaign idea, a positioning question — I open the right project and Claude already has all the context it needs.
But it's not just for work. You could create a project for a holiday — drop in your hotel bookings, restaurant reservations, a packing list, the itinerary — and then ask Claude to help you plan each day, check whether you've forgotten anything, or find things to do near wherever you're staying. The documents travel with the project, so everything's in one place rather than scattered across your inbox and a dozen browser tabs.
It connects to your actual life
Claude now connects directly to Google Workspace — Gmail, Google Calendar, Google Drive — without any complicated setup. You authorise it once and it can actually read and work with your real information.
Ask it "what have I got on today and are there any emails I haven't replied to?" and it'll go and check. Use it as a morning briefing. Ask it to draft a reply to a specific email thread. Have it find a document in your Drive.
This sounds like a small thing, but the value is in reducing the constant tab-switching and context-assembling that makes knowledge work more tiring than it needs to be. Claude becomes the single place you think rather than another tool alongside the pile you already have.
Slack is also supported, with more integrations on the way.
Is it worth paying for?
Claude has a free tier — and critically, memory is included on the free tier now. That's a reasonable place to start and see whether it fits how you work.
Claude Pro is £18/month in the UK. ChatGPT Plus is £20/month. If you're already paying for an AI assistant and not getting persistent memory, Projects, Google integration, and genuinely better writing, you're paying similar money for a noticeably worse experience.
The paid tier gets you more usage before hitting limits, plus access to Cowork — Claude's computer-use feature, where it can actually operate your computer to complete tasks on your behalf. That one's still early but improving quickly.
How to actually switch
It takes about twenty minutes to get properly set up.
Start by telling Claude about yourself. Open a conversation and give it some context — what you do, what you're working on, how you like to communicate. It'll start building from there. Or go to Settings → Memory and paste in a brief directly.
Create a Project for your main focus. Whatever you're spending most time on right now — a work project, something personal, a big decision — make a project, drop in the relevant documents, and write a short note about what you want Claude to know going in.
Connect Google. Go to the connectors settings, link your account. Try the morning briefing. See if it's useful — most people find it is.
Give it examples of your voice. Paste in a few things you've written and ask it to match your style. Then save that as an instruction on your project. The difference in output is substantial.
ChatGPT did something genuinely important — it made AI tools approachable and mainstream. That's worth acknowledging. But it's 2026, and the tools have moved on. If you've been meaning to see what else is out there, Claude is where most people who've made the switch end up staying.
Curious about how AI tools fit into real business workflows? Get in touch — it's something I think about a lot.
