How to use Ask Anything to find behavior triggers in VillageMetrics
To find behavior triggers using Ask Anything, record daily voice journals that include context about what happened before, during, and around incidents—then ask questions like "What patterns exist around meltdowns?" or "Are meltdowns more common on certain days?" The AI analyzes everything you've mentioned across all your entries and surfaces correlations you'd never spot yourself.
Here's how to get the most out of it.
What "Ask Anything" Does
Ask Anything lets you query your journal history in plain language. Instead of scrolling through entries or building spreadsheet formulas, you ask questions like you'd ask a person:
"What patterns do you see around aggression?"
"When do meltdowns tend to happen?"
"Is there a connection between sleep and behavior the next day?"
"What was different about the good days this month?"
The AI reads through your entire journal history, looks for patterns, and gives you an answer with specific examples from your entries.
Step 1: Build a Foundation of Rich Journal Entries
Ask Anything can only find patterns in data you've captured. The richer your journal entries, the more triggers the AI can identify.
When recording, include context like:
Time of day things happened
What came before an incident (transitions, demands, sensory environments)
Who was present
How your child slept the night before
What activities or events happened that day
How incidents resolved
Example entry: "Meltdown around 4:30 when he got home from school. It was gym day, and he mentioned the gym was really loud. He'd only slept about 6 hours last night. The meltdown lasted maybe 20 minutes. He calmed down after using the weighted blanket."
This single entry gives the AI multiple potential trigger signals: after school, gym day, loud environment, poor sleep, time of day.
Step 2: Ask Open-Ended Pattern Questions
Start broad to see what patterns exist, then drill down.
Good starting questions:
"What patterns exist around meltdowns?"
"When does aggression tend to happen?"
"What do the bad days have in common?"
"What's different about good days vs. bad days?"
What you might learn:
"Meltdowns are most common between 4-5 PM, especially on school days"
"Aggression often follows mentions of poor sleep or schedule changes"
"Good days frequently mention outdoor time and consistent routine"
Step 3: Ask Specific Follow-Up Questions
Once you see general patterns, ask targeted questions:
"Are meltdowns more common on gym days?"
"How does sleep affect behavior the next day?"
"Is behavior worse when Dad is traveling?"
"Do meltdowns happen more often after school or on weekends?"
Example flow:
"What patterns exist around meltdowns?" → AI says they cluster in late afternoon
"Are afternoon meltdowns more common on school days?" → AI says yes, 3x more likely
"What happens at school on the bad days?" → AI notes gym days and mentions of substitute teachers
Step 4: Explore Correlations You Hadn't Considered
Ask Anything can surface connections you didn't think to look for:
"Are there any unexpected correlations with bad days?"
"What factors seem to predict good behavior?"
"Is there a pattern with food or meals?"
"Do any people correlate with better or worse days?"
Sometimes the AI will find things like: "Bad days often include mentions of stomach ache or physical discomfort" or "Behavior scores are higher on therapy days."
Questions That Work Well for Trigger-Finding
| What You Want to Know | How to Ask |
|---|---|
| General patterns | "What patterns exist around [behavior]?" |
| Time-based triggers | "When does [behavior] tend to happen?" |
| Day-of-week patterns | "Is [behavior] more common on certain days?" |
| Environmental triggers | "Is there a connection between [factor] and [behavior]?" |
| Comparing periods | "Was [behavior] different this month vs. last month?" |
| What helps | "What seems to help when [behavior] happens?" |
What Ask Anything Can't Do
It can't identify triggers you never mentioned. If you don't mention that it was gym day, the AI doesn't know. Richer entries = better pattern detection.
It finds correlations, not proven causes. "Meltdowns are more common on gym days" is a pattern worth investigating, but you'll need to work with your care team to confirm why and what accommodations might help.
It doesn't observe your child directly. It only knows what you tell it. If your child is reacting to something sensory that you don't notice, it won't appear in the analysis.
Frequently Asked Questions
What questions should I ask first when I'm just getting started?
Start broad: "What patterns exist around meltdowns?" or "What do the bad days have in common?" Then drill down based on what surfaces. If the AI says meltdowns cluster in late afternoon, follow up with "Are afternoon meltdowns more common on school days?" You're building a hypothesis, then testing it.
What if Ask Anything doesn't find any patterns?
It might mean there isn't an obvious pattern yet, or you need more data. It could also mean your entries aren't including the relevant context. Try adding more detail about what happens before incidents, sleep, and environment—then ask again in a couple weeks.
Can I ask about patterns for a specific time period?
Yes. You can ask "What patterns existed in November?" or "Compare triggers this month vs. last month." This is useful for seeing if patterns change over time or correlate with medication changes or life events.
Ready to stop flying blind? VillageMetrics turns your daily voice notes into the data doctors need to help your child.