Getting Better Results from AI: 3 Common Mistakes to Avoid

AI is a powerful tool for making community data easier to use, but it’s also new and imperfect. It takes a little practice — as it would with a human assistant! — to figure out how to communicate with (or prompt) Sidekick to get the results you want, and often, you have to learn from your mistakes. 

While we continue to improve Sidekick's abilities, check out these 3 common mistakes people make when asking for help. And don't worry, we've included easy ways for you to learn how to get what you want with a few small tweaks!

1. More maps, more problems

The first, most common mistake we've seen is mapping overload. When Sidekick suggests several datasets for your query about housing affordability, it’s very easy to get excited and say, “Yes, map all of this!”

But, maps can only contain 1 or 2 variables, so be specific about what you want.

Avoid: “Create a map of renters by race/ethnicity.”
Try: “Create a map of Hispanic renters.”

Pro Tip: If you want to map lots of data, create multiple maps…one at a time. 

2. Google vs. Sidekick

Ah-hem, treat Sidekick like a smart assistant, not a search engine. You wouldn’t turn to another person and blurt out “children” or “poverty." Instead, you’d offer some context to guide your assistant’s search and help ensure its success. The more context you give Sidekick about what you're looking for, the better it can understand and return a more favorable and helpful result.

Avoid: Single-word searches like “poverty”
Try: “What data should I use to examine housing and poverty?”

3. Not all at once, please!

No one likes to feel like they're drinking from a firehose, including Sidekick. A future where AI can instantly, intelligently answer complex, multi-step questions is in sight — but it’s distant. For now, avoid cramming all your questions into one query. Start with one request, and build from there. For instance, ask about food insecurity data first, then request a visualization. 

• “What data on employment could inform a workforce development plan to help residents find living wage jobs?”
• “What data on education could inform a workforce development plan to help residents find living wage jobs?”
• “Looking at this data, identify the top 2-3 key takeaways for community stakeholders.”
• Map {specific variable - see point above!} at the most granular level possible.”

An assistant that works smart and hard 

The more us humans hone our skills working with AI, the more we can get out of the technology. Sidekick can save teams immense amounts of time and energy finding data, uncovering insights, creating interactive visualizations, and tailoring narratives for any audience or community — asking good questions is key to getting started.

Want to read more about how AI can be used for good? Hear from CEO Stephen Hardy asking great questions can make all the difference on your journey to making impactful change. Read the blog →