I’m a Notion power user. My entire professional life lives in Notion—project notes, meeting records, reading lists, documentation, team wikis. When Notion AI launched, I adopted it immediately. It was helpful, but after months of use, I started hitting its limits.
Now I’ve found a better approach: using Notion AI and ChatGPT together, each for what it does best.
## What Notion AI Does Well
Notion AI isn’t trying to be a general-purpose AI assistant. It’s a document-native tool, designed to work within your existing Notion workspace. That focus creates specific advantages.
**Inline editing and refinement**
This is Notion AI’s killer feature, in my view. When I’m writing in Notion and want to improve a paragraph, I just select the text, click “Ask AI,” and choose an action—improve writing, make shorter, fix spelling, change tone to professional, whatever I need.
The workflow is seamless. I don’t have to copy text, switch to another app, generate output, and paste it back. The AI operates directly on my document.
My common use cases:
– Grammar and spelling checks as I write (so much faster than waiting until the end)
– Condensing verbose paragraphs into clear sentences
– Translating English text to Chinese or vice versa
– Generating a quick summary of a long section before sharing
**Organizing messy notes**
I take notes quickly—often in fragments, in shorthand, with abbreviations I’ll only understand for a few days. When I need to use these notes later, I paste them into Notion AI and ask it to organize into a structured format.
Example: My raw note from a client call reads:
“met wang today, he said proposal is good but budget needs review. tech team says 2 more weeks. also wants to see first version before end of next week. asked about support after launch.”
Notion AI transforms this into:
– **Discussion Points:** Proposal approval (conditional on budget review), timeline extension of 2 weeks requested by tech team
– **Client Requests:** First version demo before end of next week; clarification on post-launch support
– **Action Items:** Review budget constraints with finance; coordinate with tech team on new timeline
Much easier to work with when I’m writing the follow-up email or project plan.
**Quick translations within documents**
When I need to translate a section for a bilingual colleague, Notion AI handles it inline. Select, translate, done. No switching tabs or tools.
## Where Notion AI Falls Short
Despite its strengths, Notion AI has real limitations that became frustrating over time.
**It can’t handle complex, multi-step tasks**
Notion AI is designed for quick document operations—edits, summaries, translations. It doesn’t handle complex analytical tasks well at all. Asking it to write a comprehensive market analysis or debug a piece of code produces weak results. It was never designed for this, and it shows.
**No real-time information**
Notion AI doesn’t know what’s happening today. Ask it about recent news, current market data, or the latest research, and you get outdated information or an admission it doesn’t know.
For anything requiring current information, I have to leave Notion AI entirely.
**No conversational context**
Each Notion AI interaction is isolated. There’s no memory of previous queries, no ability to build on earlier responses. This makes it unsuitable for anything that requires iterative refinement or back-and-forth.
**Chinese language understanding isn’t perfect**
Notion AI sometimes struggles with context-specific Chinese meanings. It might not understand industry jargon or context-dependent phrases the way a more sophisticated model would.
## What ChatGPT Does Better (For My Needs)
**Complex analytical tasks**
When I need to write a comprehensive analysis, competitive research, or detailed plan, ChatGPT is my tool. I paste in context, ask follow-up questions, build iteratively, and get substantive output.
**Creative and flexible tasks**
For brainstorming, ideation, and exploring different approaches, ChatGPT feels more capable of genuine creativity and flexibility. Notion AI produces solid output; ChatGPT can surprise you.
**Current information**
ChatGPT with browsing can search the web and synthesize current information. Notion AI can’t do this at all.
**Multi-turn conversations**
I can build on previous responses, clarify, refine, and develop ideas over multiple turns. The conversation has memory and context in a way Notion AI doesn’t.
## My Current Combined Workflow
After experimenting for a few weeks, here’s how I’ve integrated both tools:
**For writing and note-taking in Notion:**
– Draft in Notion
– Use Notion AI for quick edits, grammar checks, and translations
– When I need deeper content development, copy relevant text to ChatGPT
**For complex tasks:**
– Start in ChatGPT
– Develop the core content through conversation
– Paste refined output back into Notion
**For research:**
– Use ChatGPT with browsing for initial research
– Organize findings in Notion
– Use Notion AI to structure and format
The combination is more powerful than either tool alone. Notion AI’s seamless document integration makes it essential for writing workflow. ChatGPT’s capabilities make it essential for anything requiring deep thinking or current information.
## Does This Justify Both Subscriptions?
Notion AI comes with Notion Plus at $10/month. ChatGPT Plus is $20/month. Combined, that’s $30/month for AI tools.
Is it worth it? For me, yes. The productivity gains in writing speed and research efficiency probably save me 5-10 hours per month. At market rates for my time, that’s easily worth the investment.
Your calculation might be different. If you’re a casual user, the free tiers of both tools might be sufficient. If you live in Notion and write extensively, Notion AI is probably worth the $10. If you need powerful AI for complex tasks, ChatGPT Plus is probably worth $20.
Both together only makes sense if you’re using both extensively. If you’re not sure, start with one and add the other if you find yourself needing capabilities it doesn’t have.
## The Key Insight
The real lesson from this experiment is that no single AI tool is optimal for everything. Different tools have different strengths, and combining them strategically leads to better results than trying to force one tool to do everything.
Notion AI is excellent at what it does—document-native AI assistance. ChatGPT is excellent at what it does—general-purpose AI assistance. Together, they cover more ground than either alone.
That’s been my experience. Questions or different approaches? I’d love to hear how others are combining AI tools in their workflows.
Neil Shum is a 10-year internet industry veteran with experience spanning product management, startup founding, and AI-native product development.
Starting his career at a Fortune 500 tech company, Neil spent his early years deep in product strategy and user research. In 2018, he co-founded a H5 game startup that scaled to 500,000 users before being acquired in 2022.
These days, Neil focuses on exploring how AI is reshaping product design, user experience, and business models. He’s particularly interested in the practical side of AI adoption—what works, what doesn’t, and what founders and product teams should actually pay attention to.
When not analyzing AI tools or writing about emerging trends, you’ll find him testing new AI products, mentoringearly-stage founders, or reading way too many newsletters about LLMs.