Starting Small

Theory is useful. But at some point, your team needs to actually do something. Here are specific tasks that score low across all four risk dimensions — high context, focused output, read-only, human-initiated.

Editing and Formatting

These are nearly risk-free because the content already exists. AI is just improving it. High context-to-output ratio by definition.

Grammar and clarity editing. Paste a human-written draft and ask AI to improve grammar, tighten language, or restructure for clarity. Compare the AI version against yours. Accept what’s better, reject what isn’t.

Reformatting. Take a narrative and ask AI to restructure it into a different format — bullet points, a table, an executive summary. The content is yours; the AI handles the layout.

Tone adjustment. Ask AI to make a section more formal for a federal funder, or more conversational for a family foundation. Compare the results against your sense of what’s appropriate.

Template generation. Ask AI to create empty templates — budget shells, reporting frameworks, meeting agendas — based on your specifications. You fill in the real content.

Research and Summarization

These tasks involve AI processing existing information. The output is focused, and you can verify against the source material.

RFP summarization. Upload a long RFP and ask AI to extract the key requirements, deadlines, eligibility criteria, and evaluation criteria. Cross-check against the original document.

Funder profile summaries. Give AI a funder’s website URL or 990 data and ask for a summary of their priorities, giving history, and typical grant sizes. Verify key facts.

Document comparison. Ask AI to compare two drafts of a proposal section and highlight the differences. Or compare your draft against the RFP requirements.

Meeting notes and follow-ups. After a funder meeting, give AI your rough notes and ask it to generate organized minutes and action items.

What to Watch For

Even with low-risk tasks, pay attention to:

AI’s formatting preferences may not match yours. AI loves bullet points and headers. Your organization may prefer narrative paragraphs. Don’t let AI’s default style creep into your work uncritically.

Summaries can miss nuance. When AI summarizes an RFP, it may deprioritize a buried requirement that’s actually critical. Always check the summary against the full document.

Tone adjustments may lose substance. When AI makes something “more concise,” it sometimes cuts content that matters. Review every change, not just the overall impression.

Making It a Team Activity

The best way to start small is to start together:

Pick three tasks from the lists above that your team does regularly
Have everyone try them during the Science Fair week or a designated practice period
Compare results — did AI save time? Was the output usable? What needed fixing?
Keep what works — add the useful tasks to your regular workflow

The goal isn’t to use AI for everything. It’s to find the two or three tasks where AI genuinely helps and make those a habit.

Check your understanding

Your team is starting with AI for the first time. Which task should they try first?

Key Takeaways

Start with editing, formatting, and summarization -- tasks where the content already exists
Cross-check all AI research and summaries against source material
Watch for style creep, missing nuance, and substance loss in edits
Pick three low-risk tasks, try them as a team, and keep what works

### Next Lesson

You’ve started small. Now let’s plan how to expand — taking on tasks that score higher on the risk dimensions as your team’s confidence and skills grow.