I use AI every single day. But for a long time, I was using it wrong. Like most people, I defaulted to one chatbot for everything just because it felt silly to change tabs. What could go wrong with using the same chatbot for brainstorming, research, drafting outlines, editing or decision-making? As it turns out, quite a lot.
It wasn’t obvious at the time, but some outputs felt shallow or way too verbose. And sometimes, I’d get stuck in a loop of “almost good enough” just to rework everything on my own. The point of AI is to reduce workflow, not cause more work.
So, I stopped treating AI like a single assistant — and started using more chatbots as a way to beef up my “team.” Now, I use Claude, Gemini and ChatGPT for completely different roles. The result is better ideas, much faster research and far less bottlenecks so I actually get work done.
Here’s exactly how my system works.
Claude is my strategist
(Image credit: Future/NPowell)
When I need to really think on a project, I go to Claude. Kind of like phoning a friend you know you can count on to give you the best feedback, that’s the way I see Anthropic’s flagship model.
Article continues below
You may like
Out of all the models I’ve tested, Claude consistently feels the most structured and deliberate. It doesn’t just generate ideas — it organizes them. It also doesn’t love all of my ideas or flatter me like the people-pleasing ChatGPT.
I use Claude for stress-testing arguments, identifying gaps in logic, turning messy thoughts into a clear direction and framing angles for articles and projects.
One of my go-to prompts is: “What’s weak, unclear or missing from this idea?”
Claude is especially good at pushing back. It will tell you when something is too obvious, too vague or missing a strong takeaway — which is exactly what you want at the strategy stage.
Why this works: Claude shines at top-down thinking. It helps you figure out what you should say before you worry about how to say it.
Gemini is my researcher
(Image credit: Future)
When I need to learn fast, I switch to Gemini. Gemini is incredibly good at pulling together structured information, especially when the topic is complex or evolving. This is especially true with all of Google AI Labs tools with Gemini. NotebookLM is my go-to for all things research and learning. I even use the app to learn on the go.
I use Gemini every day to help make sense of complex news stories, summarizing long reports, comparing tools and models or simply pulling key facts into clean formats.
What to read next
One of my favorite Gemini prompts is: “Summarize this for a general audience. What are the 3 most important takeaways and why do they matter?”
Gemini tends to be more factual and grounded, which makes it ideal for research-heavy workflows.
Why this works: Google’s flagship model is smart, fast and structured so it’s less likely to drift into unnecessary creativity or fluff when you need clarity.
ChatGPT handles the grind
(Image credit: Gabby Jones/Bloomberg/Getty)
Once I know where my project is headed and have enough information researched, I move on to ChatGPT for the rest. OpenAI’s model has become my best bet for execution. It takes care of the time consuming parts of the job.
I use it as a thesaurus to help rewrite with clarity, formating outlines and handling day-to-day tasks like chatting live, fact-checking or scam detection.
A prompt I often use is: “Take what I give you for this project and tell me what I need to do to move this forward.”
This is where ChatGPT really shines — speed and adaptability. It’s also incredibly good at iteration. Once the thinking is done, it helps you move faster without losing quality.
The system that changed everything
Here’s what this looks like in practice:
Start with Claude: Define the idea and angleMove to Gemini: Gather facts and structure researchFinish with ChatGPT: Execute, refine and publish
By using this system I count on each different tool to focus on what it does best. If you’re relying on one tool for everything, using AI might actually be holding you back.
Keep in mind that it’s simply not supposed to be a one-size-fits all plan. Utilizing AI as a stack and assigning each model a role means you get better outputs with fewer prompts and spend less time fixing mediocre results. It’s not about better prompting. It’s about better systems.
Bottom line
If you’re only using one AI tool, you’re leaving a lot on the table. Splitting your workflow across Claude, Gemini and ChatGPT doesn’t make things more complicated — it actually helps reduce bottlenecks so things run smoother. So give it a try. Let Claude help you think through your project like a confident guide, let Gemini help you maximize research and then give ChatGPT everything to help you execute your project. Once you start using AI this way, it’s very hard to go back.
Follow Tom’s Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom’s Guide