Tips & Best Practices

Getting better results

Complete your organization profile

The AI uses your organization profile as context for every response. A thorough profile — mission statement, programs, geographic focus, financials, theory of change — dramatically improves the relevance and accuracy of AI output across the entire platform. Use the /profile skill to build or update it.

Upload source materials early

Past successful proposals, annual reports, program descriptions, and evaluation reports all become reference material the AI can cite. The more relevant material in your file tree, the more evidence-based the AI’s output becomes.

Be specific in your prompts

“Draft the needs statement for the youth mentoring program using data from the 2025 annual report” works much better than “Write something about our programs.” The more specific your request, the more targeted the AI’s response.

Use skills for focused work

When you have a clear task — drafting, prospecting, reviewing — invoke the matching skill with a slash command. Skills give the AI a structured workflow tuned for that task. See Skills & Slash Commands for the full list.

Choose the right model tier

Use auto for most tasks, pro for complex writing that needs the highest quality, and fast for quick questions and simple lookups. See AI Model Tiers for details.

Understanding AI output

Citations and evidence

When the AI references source materials, it cites specific documents. Check these citations — they tell you exactly where the AI found its information. If a citation seems wrong or the AI couldn’t find evidence, that’s a signal to provide better source materials.

Confidence signals

The AI communicates its confidence level. High-confidence responses are backed by strong evidence in your materials. When the AI hedges or says it couldn’t find information, it’s being honest — not failing. Use those moments to fill gaps in your uploaded content.

When the AI asks for input

Sometimes the AI determines it can’t complete a task without information from you — budget numbers, specific program details, or strategic decisions that aren’t in your documents. Providing this information directly in the chat gives the AI what it needs to continue.

Common patterns

For new applications

1. Upload the RFP and any relevant source materials
2. Ask the AI to analyze the requirements and outline an approach
3. Use /write to draft sections based on the outline
4. Use /review to check the draft against RFP requirements

For evaluating opportunities

1. Ask the AI to assess your fit with a funder or opportunity
2. Review the fit score and evidence for each criterion
3. Use the results to decide whether to invest application effort

For improving existing drafts

1. Attach your draft to a message
2. Ask the AI to review it against specific requirements, or use /review
3. Use follow-up prompts to revise specific sections

What the AI won’t do

• Fabricate information — The AI won’t invent statistics, program outcomes, or organizational details. If it can’t find evidence, it says so.
• Override your decisions — The AI proposes and recommends, but every action requires your confirmation.
• Work without context — The AI’s quality depends on the materials you provide. Generic prompts without context produce generic output.