Gemini quickly became a core part of my workflow after Google launched it as Bard.
Google’s deep Gemini integration across its products and services only increased my reliance on its AI chatbot and tools.
However, after months of dealing with some frustrating Gemini limitations, I switched to Anthropic’s Claude.
That was three months ago. Claude has changed the way I work with AI, enabling me to get more done without running into weird bugs or issues.
There’s no going back to Gemini now.
I thought Gemini was enough until I finally tried Claude
I ignored Claude for months, and now I get the hype
Gemini was useful until I started fact-checking its answers
Gemini’s inaccuracies made it hard to trust for research
I don’t use AI for coding or anything overly technical. Instead, my workflow revolves around research, writing, creating reports, brainstorming, and summarizing lengthy documents.
Gemini worked well for these tasks, with its Google integration only helping my workflow.
I could ask Gemini to surface an email from my inbox and draft an appropriate response. Or I could ask it to pull my hotel booking details and flight details and then create a stroller-friendly holiday itinerary.
But the more I used Gemini, the more I realized how badly it fumbled in some cases.
When using Gemini for research purposes, I noticed it would confidently cite made-up sources or give incorrect information. This has happened multiple times over the last few months.
While hallucinations are a common issue with AI chatbots, Gemini’s confidence in its responses was concerning.
It could have told me that it was unsure about the source of the information or the accuracy of the information.
If anything, this made me skeptical about relying on Gemini’s responses for any serious work. Since I relied heavily on Gemini for research and brainstorming, this was a major roadblock.
Gemini doesn’t connect with the apps I actually use
Limited third-party app support kept disrupting my workflow
As much as I find Gemini’s integration with other Google services useful, it’s also a limiting factor. Outside of Google apps, Gemini only works with a handful of apps and services, like Spotify.
Google only recently got around to adding Canva integration. Yet, the list of supported third-party apps and services is still limited.
While Google Tasks is great, I prefer using Todoist. And since Gemini can’t talk to Todoist, I have to manually add important tasks to my to-do list.
Likewise, I use ClickUp for project management, which Gemini does not integrate with. So, every time I needed to update a project status based on my research, I manually switched to ClickUp to do it.
Why Claude won me over
It’s accurate and works with more apps
After being frustrated with these Gemini limitations, I gave Anthropic’s Claude a shot.
Based on all the chatter in my circle, I assumed Claude was only great for coding and technical tasks. I couldn’t be more wrong.
Claude’s responses are noticeably more accurate and consistent. More importantly, it doesn’t hallucinate nearly as much as Gemini.
If Claude doesn’t know the answer to a question, it will tell that straight up rather than fabricate a response.
This may sound like a minor difference, but for the kind of work I do, it’s critical.
I regularly use AI to research income tax laws and help stay on top of some ongoing court cases for a friend.
Gemini frequently gave straight-up wrong information or cited made-up sources. Claude has yet to falter, always giving detailed and useful insights.
What has made Claude an indispensable tool in my workflow is its ability to integrate with other apps and services.
Thanks to MCP (Model Context Protocol) support, Claude can talk to a plethora of third-party apps.
After I’m done researching a project, I ask Claude to add the actionable tasks to Todoist. Or, when working on a work-related project, I can tell Claude to give an overview of the work done by my team by pulling the relevant data from ClickUp.
MCP is an open protocol, and it enables Claude to do a lot more. Gemini can only provide instructions on what to do in an app or service.
Thanks to MCP support, Claude can actually take the steps for me, getting stuff done faster and more efficiently.
Gemini supports MCP, but requires using the Gemini CLI (Command Line Interface). By comparison, I can paste an MCP’s URL inside a Claude chat box and ask it to do the rest.
Claude Cowork is also another great feature. I can point it to a folder, dump all project-related documents into it, and ask Claude to create actionable tasks and reports based on the documents.
This kind of work used to take up hours of my work. Now, thanks to Claude, I can get it done in minutes.
Claude isn’t perfect, but I’m not going back
Claude is not without its flaws.
Coming from Gemini, I’ve frequently run into usage limits. I hit them more often than I like, especially on busy workdays.
Plus, within Google’s ecosystem, Gemini is still hard to beat. It’s available across all services that I use, such as Gmail, Docs, and Sheets.
That’s a convenience Claude can’t match. However, convenience can only get you so far.
Claude has won me over with its more reliable responses and the ability to work with a lot more third-party apps.
I can use Claude for more serious work, knowing that it won’t confidently give false answers.
On top of that, it connects and works with tools that I use every day, helping me work more efficiently.
