Projects vs Custom AI Assistants: How to Pick the Right Tool

🧭 THIS WEEK AT AI SECOND ACT

Howdy!

I've been watching smart professionals wrestle with the same question:

"Should I build a custom chat for this, or just use a project?"

Most are doing it backwards. Building complex custom GPTs for simple tasks. Using basic ChatGPT for organized work that needs structure.

Today, I'm giving you the decision framework that'll save you hours of confusion.

If you’re enjoying this newsletter, please share it! Forward this to a friend/colleague and have them subscribe at aisecondact.com/subscribe. Thanks!

👉 You can also just hit "Reply" and let me know what you want more (or less) of. There is also a poll at the end to give feedback.

My goal is to make this as valuable and practical as possible as we navigate the new AI era. 🚀

🧰 AI NEWS + LEARNING

Here’s a few things I found recently:

• Wispr Flow —> Effortless Voice Dictation

• Team AI —> Access multiple chat tools from one UI

• Standford 2025 AI Report

• Manus Slides —> Manus is AMAZING - way more than a chatbot, it’s an autonomous agent capable of full task execution! - I need to dedicate a full edition to this soon.

🗺️ FEATURED INSIGHT

Projects vs Custom AI Assistants: The Decision Guide You Need

Here's the thing: these aren't competing tools. They're different solutions for different problems.

Remember, my favorite chatbots:

• ChatGPT - general all-round use, strong models, multi-modal support [text, image, video], web search, research, strong reasoning, external connections, best canvas feature

• Perplexity - best research tool, amazing labs feature

• Claude - best writing tool, research, reasoning, external connections, canvas

Think of it this way:

• Projects = Your personal AI filing cabinet

• Custom AI Assistants = Your specialized AI employees

What Are AI Projects?

Projects are your organized workspace. You can:

• Upload files once and reference them across multiple conversations

• Keep different topics in separate spaces (work, personal, specific campaigns)

• Maintain context across weeks or months

• Set custom instructions or goals for each project space

Key Insight: Projects are about organization and persistence for your work across varied contexts. They're ideal for things like drafting long-form documents over time, collecting and refining expert insights, curating ongoing research, or managing evolving client work across multiple inputs.

What Are Custom AI Assistants?

Custom AI Assistants—such as ChatGPT's Custom GPTs or Gemini's Gems—are pre-configured AI chatbots with:

• Built-in expertise and custom instructions

• Optional knowledge bases or datasets

• Consistent output styles

• Shareable links (so others can use them too)

• Separate conversation history per user

Key Insight: Custom AI Assistants are about specialization and shareability. They are created for a specific, typically narrow-focus purpose—like generating weekly marketing status reports, summarising client interviews using a specific framework or outputting formatted meeting minutes.

The Decision Framework

Use Projects when:

✅ You're managing ongoing personal work
✅ You need file persistence across multiple sessions
✅ You want organized spaces for different topics
✅ You're doing research that spans weeks/months
✅ You need different contexts for different types of work

Example: You're working on a quarterly strategy presentation. Upload all your research docs to a "Q3 Strategy" project. Every conversation builds on the previous context. Files stay accessible.

Build Custom AI Assistants for:

✅ Processes you want to share with your team
✅ Specialized knowledge that others need access to
✅ Consistent outputs (like status reports or analysis)
✅ Company-specific tasks with unique requirements
✅ Tools that multiple people will use independently

Example: You create a "Product Requirements Analyst" assistant loaded with your company's PRD templates and standards. Share the link with your team. Everyone gets consistent, on-brand outputs.

The 80/20 Rule for Projects vs Custom Chats

80% of your AI workflow problems are solved by Projects.

They handle the messy reality of how we work:

• Jumping between conversations

• Building on previous research

• Keeping different workstreams separate

• File management & knowledge

Custom AI Assistants solve the other 20% – when you need something specialized or shareable.

Real-World Examples

Projects Setup:

• "AI Second Act" (newsletter research, drafts, templates)

• "Hobbies" (running & astrophotography, training logs, photo editing workflows, research notes)

• "Learning & Development" (course materials, tutorial research)

Custom Assistant I Built:

• "Program Management Assistant" (loaded with PM frameworks, risk templates, and status report formats)

• Shared with the team

• Everyone gets consistent project outputs

The Quick Decision Test

Ask yourself:

1. Will others need this? → Custom Assistant

2. Is this ongoing work? → Project

3. Do I need file persistence & knowledge? → Project

4. Does this require specific expertise? → Custom Assistant

5. Am I organizing vs. specializing? → Project vs. Custom Assistant

Tool Availability Check

Not all AI tools offer both features. Here's what's available:

Getting Started This Week

If you've never used Projects:

1. Open your AI tool of choice (ChatGPT, Gemini, etc.)

2. Create your first project: "Work Admin"

3. Upload a few key documents if any

4. Set custom instructions: "I'm a [your role] who needs help with [your typical tasks]"

5. Start all conversations there

If you need your first Custom Assistant:

1. Think of one process your team does repeatedly

2. Identify what knowledge/templates it needs

3. Build it with clear, specific instructions

4. Test it yourself first, then share the link

[note, btw I did not add specific instructions here - partly because it’s tool dependent, it may get outdated soon and of course, it becomes a much longer read. Reply and let me know if you want more detail in the future, even perhaps just on the website itself].

Common Mistakes to Avoid

❌ Building a custom assistant for personal file management
❌ Using basic AI chat for team processes
❌ Making custom assistants too complex (start simple)
❌ Not setting up Projects with custom instructions
❌ Forgetting that Projects are private, Custom Assistants can be shared

The Bottom Line

Most people need Projects first. They solve the "where did I put that conversation?" problem and keep your AI work organized.

Custom Assistants come later when you've identified processes worth systemizing and sharing.

Start with Projects this week. Your future self will thank you.

👉 Hit "Reply" and tell me: What's your biggest AI organization challenge right now? I read every response and often turn them into future newsletters.

P.S. If this helped clarify the Projects vs Custom chat confusion, please forward it to a colleague who's wrestling with the same decision. These tools are most powerful when teams understand how to use them properly.

Future-proof your career with AI
— Brett

_{Picture by Octavian-Dan on Unsplash.}