
After playing around with ChatGPT or Claude for a while, you’ve probably thought, “This is helpful, but I wish it knew more about our work.” That’s where custom AI tools come in! Let me share how I create specialized AI assistants that boost the effectiveness of my work.
Why Generic AI Just Isn’t Enough
Generic AI tools are great starting points, but they have limitations:
- They don’t know your organization’s history, programs, or voice.
- They can’t access your grant history, program data, or impact metrics.
- They don’t understand your sector’s unique terminology and challenges.
Custom AI tools tackle these challenges by creating a dedicated assistant that embodies your nonprofit uniquely. It’s like having a team member who has examined every document you’ve ever produced.
What Are Custom GPTs and Research Vaults?
A custom GPT (or “Brain”) allows you to develop persistent instructions and a consistent set of trusted data resources. Think of it as your private AI assistant, trained on your organization’s materials.
I’ve created several custom tools that save me hours each week:
- Susan’s Brain: An AI Tool that Writes in My Voice
- Usage: Edits blog posts and newsletters and creates metadata and social media content that sounds exactly like me
- AI Acceptable Use Policy Ninja: Helps craft acceptable use policies based on best practices
- Usage: Supports nonprofit clients who need guidance on responsible AI implementation
These custom GPTs are focused, quick, and less likely to make mistakes because I defined the purpose and the dataset. They are enhanced versions of AI that retain knowledge about specific topics.
How I Use Custom AI Daily
Here’s how these tools transform my workflow:
- Research, analyze, and edit grants
- My custom GPT research tools, which I use for clients like Northern California College Promise Coalition, share data, analyze RFPs, and edit drafts.
- Review and summarize reports.
- I upload PDFs of reports and then ask for relevant summaries and great quotes to save hours of reading time.
- Follow up with clients and funders.
- I use AI to refine my meeting notes and draft follow-up emails that maintain relationships.
The best part? After you set up these tools, they become increasingly valuable as you add more information..
Building Your Own Custom Brain (It’s Easier Than You Think!)
Creating a custom brain or research vault requires just three things:
- A premium account at ChatGPT or Claude ($20/month).
- A set of instructions for what you want the tool to do/
- A set of materials—PDFs preferred—to use as the primary reference set.
The process is surprisingly straightforward:
- Gather your materials: Gather program descriptions, grant narratives, annual reports, style guides, and other documents that reflect your organization’s work and voice.
- Provide clear instructions: Communicate what you want the AI to do. Be specific about tone, limitations, and purpose.
- Upload and organize: Add your documents to the custom GPT or Claude project.
- Test and refine: Ask questions and evaluate responses, adjusting your instructions until you receive consistently useful results.
I’ve created step-by-step guides that walk you through this process:
- Empower Your Grant Writing with An AI-Powered Research Brain
- DIY Custom AI: A Step-by-Step Guide to Building Your Own GPT
Real-World Examples That Make a Difference
Let me share some actual screenshots of my custom tools:
These aren’t just theoretical examples—these are tools I use daily that save me hours of work while improving quality and consistency.
Where Custom AI Delivers the Biggest ROI
These organizational areas each benefit from custom AI tools:
- Development: Custom GPTs that understand your grant history, funding priorities, and past successes can dramatically speed up proposal development.
- Communications: An AI trained on your style guide and past content can draft consistent messaging across all channels.
- Program evaluation: Tools that analyze your impact data can help identify trends and generate compelling outcome narratives.
- Board management: Create a board-specific assistant that knows your governance documents, strategic plan, and meeting protocols.
The time savings are immediate, but the real value is consistency. Your custom AI maintains and consistently applies institutional knowledge; you can add new materials as needed.
Your Next Step: Identify Your First Custom AI Opportunity
Look at your workflow and ask: What specialized knowledge do I repeatedly need to access? What documents do I constantly reference? Where am I repeatedly recreating similar content?
That’s your opportunity for a custom AI solution.
Start small—perhaps a custom GPT that knows your messaging guidelines or grant narratives—and expand as you see results. The investment is minimal, and the potential return is enormous.
Do you have questions about building your custom AI tool? Are you wondering which of your materials would be most valuable to include? Drop me a line. I’d love to hear what you’re working on and how I might help!
This post is part of a four-part series. See all the posts here:
After playing around with ChatGPT or Claude for a while, you’ve probably thought, “This is helpful, but I wish it knew more about our work.” That’s where custom AI tools come in! Let me share how I create specialized AI assistants that boost the effectiveness of my work.
Why Generic AI Just Isn’t Enough
Generic AI tools are great starting points, but they have limitations:
- They don’t know your organization’s history, programs, or voice.
- They can’t access your grant history, program data, or impact metrics.
- They don’t understand your sector’s unique terminology and challenges.
Custom AI tools tackle these challenges by creating a dedicated assistant that embodies your nonprofit uniquely. It’s like having a team member who has examined every document you’ve ever produced.
What Are Custom GPTs and Research Vaults?
A custom GPT (or “Brain”) allows you to develop persistent instructions and a consistent set of trusted data resources. Think of it as your private AI assistant, trained on your organization’s materials.
I’ve created several custom tools that save me hours each week:
- Susan’s Brain: An AI Tool that Writes in My Voice
- Usage: Edits blog posts and newsletters and creates metadata and social media content that sounds exactly like me
- AI Acceptable Use Policy Ninja: Helps craft acceptable use policies based on best practices
- Usage: Supports nonprofit clients who need guidance on responsible AI implementation
These custom GPTs are focused, quick, and less likely to make mistakes because I defined the purpose and the dataset. They are enhanced versions of AI that retain knowledge about specific topics.
How I Use Custom AI Daily
Here’s how these tools transform my workflow:
- Research, analyze, and edit grants
- My custom GPT research tools, which I use for clients like Northern California College Promise Coalition, share data, analyze RFPs, and edit drafts.
- Review and summarize reports.
- I upload PDFs of reports and then ask for relevant summaries and great quotes to save hours of reading time.
- Follow up with clients and funders.
- I use AI to refine my meeting notes and draft follow-up emails that maintain relationships.
The best part? After you set up these tools, they become increasingly valuable as you add more information..
Building Your Own Custom Brain (It’s Easier Than You Think!)
Creating a custom brain or research vault requires just three things:
- A premium account at ChatGPT or Claude ($20/month).
- A set of instructions for what you want the tool to do/
- A set of materials—PDFs preferred—to use as the primary reference set.
The process is surprisingly straightforward:
- Gather your materials: Gather program descriptions, grant narratives, annual reports, style guides, and other documents that reflect your organization’s work and voice.
- Provide clear instructions: Communicate what you want the AI to do. Be specific about tone, limitations, and purpose.
- Upload and organize: Add your documents to the custom GPT or Claude project.
- Test and refine: Ask questions and evaluate responses, adjusting your instructions until you receive consistently useful results.
I’ve created step-by-step guides that walk you through this process:
- Empower Your Grant Writing with An AI-Powered Research Brain
- DIY Custom AI: A Step-by-Step Guide to Building Your Own GPT
Real-World Examples That Make a Difference
Let me share some actual screenshots of my custom tools:
These aren’t just theoretical examples—these are tools I use daily that save me hours of work while improving quality and consistency.
Where Custom AI Delivers the Biggest ROI
These organizational areas each benefit from custom AI tools:
- Development: Custom GPTs that understand your grant history, funding priorities, and past successes can dramatically speed up proposal development.
- Communications: An AI trained on your style guide and past content can draft consistent messaging across all channels.
- Program evaluation: Tools that analyze your impact data can help identify trends and generate compelling outcome narratives.
- Board management: Create a board-specific assistant that knows your governance documents, strategic plan, and meeting protocols.
The time savings are immediate, but the real value is consistency. Your custom AI maintains and consistently applies institutional knowledge; you can add new materials as needed.
Your Next Step: Identify Your First Custom AI Opportunity
Look at your workflow and ask: What specialized knowledge do I repeatedly need to access? What documents do I constantly reference? Where am I repeatedly recreating similar content?
That’s your opportunity for a custom AI solution.
Start small—perhaps a custom GPT that knows your messaging guidelines or grant narratives—and expand as you see results. The investment is minimal, and the potential return is enormous.
Do you have questions about building your custom AI tool? Are you wondering which of your materials would be most valuable to include? Drop me a line. I’d love to hear what you’re working on and how I might help!
This post is part of a four-part series. See all the posts here:

