Intelligence: Using AI Chat
The Intelligence Dashboard includes an AI-powered chatbot that lets you ask questions about your data in plain English. No technical knowledge required—just ask questions the way you'd ask a colleague.
Getting Started
Where to Find It
Look for the AI Chat icon in the bottom right of the dashboard interface. Click to open the chat panel, then type your question.
What It Does
The chatbot reads your portfolio data and translates your questions into data queries. It can:
Pull metrics and KPIs
Filter by buildings, tenants, or time periods
Compare performance across dimensions
Surface insights you might not have thought to look for
What It Doesn't Do
The chatbot works with data in the Intelligence Dashboard. It can't access external systems, make changes to your content, or take actions on your behalf—it's purely for analysis and exploration.
How to Ask Good Questions
Be Specific
The chatbot works best with clear, focused questions.
Instead of... | Try... |
|---|---|
"Show me data" | "Show me content views by topic for the past 30 days" |
"How are we doing?" | "Which buildings have the highest engagement rates this quarter?" |
"Tell me about tenants" | "Which tenants have the lowest action rates?" |
Use Natural Language
You don't need to know database terms or technical jargon. Ask questions the way you'd talk to a colleague:
"Which content performed best last month?"
"How does Tuesday compare to other days for engagement?"
"Show me our top 10 most active users"
One Question at a Time
You'll get clearer results if you focus on one main question per query. Once you get an answer, follow up with additional questions to dig deeper.
Example Questions
Understanding Performance
"What's our overall reach rate for the past 90 days?"
"Which content topics get the most views?"
"Show me action rates by content type"
Comparing Time Periods
"How did content engagement compare between last month and this month?"
"Show me year-over-year growth in reach rate"
"Is our posting cadence increasing or decreasing?"
Drilling into Specifics
"What's the engagement for [Building Name]?"
"Which tenants have the highest action rates?"
"Show me the top posts from the Wellness topic"
Identifying Trends
"Show me weekly posting cadence for the past three months"
"How has reach rate changed month-over-month?"
"What time of day gets the most engagement?"
Access & Utilization
"Which day of the week has the highest building traffic?"
"Show me tenants with below-average utilization"
"How do daily scans compare to last quarter?"
Working with Results
Review Before Reacting
When the chatbot returns results, take a moment to scan the numbers. Look for:
Patterns — recurring highs or lows
Outliers — unusually high or low values
Unexpected results — things that don't match your assumptions
Ask Follow-Up Questions
See something interesting? Dig deeper:
"Why is that content type underperforming?"
"Show me the individual posts driving those numbers"
"Break that down by building"
Request Visualizations
For trends and comparisons, ask for charts:
"Show me that as a chart"
"Visualize content views over time"
"Graph the comparison"
Tips for Better Results
Tip | Why It Helps |
|---|---|
Use relative dates ("last 30 days," "this quarter") | The chatbot handles date math and timezones automatically |
Name specific buildings or tenants | Removes ambiguity and gets you precise results |
Start broad, then narrow down | Get the big picture first, then drill into details |
Ask "why" questions | The chatbot can help you explore what's driving the numbers |
Verify unexpected results | If something seems off, ask clarifying questions |
Troubleshooting
"I'm not getting the answer I expected"
Rephrase your question with more specific details
Break complex questions into smaller parts
Check that your filters (building, date range) are set correctly in the dashboard
"The chatbot doesn't understand my question"
Use simpler language
Avoid abbreviations or internal jargon
Try asking the question a different way
"Results seem wrong"
Ask the chatbot to explain how it calculated the result
Check if dashboard filters are affecting the data
Compare against the dashboard cards to verify
"I don't know what to ask"
Start with these:
"What are the key metrics I should know about?"
"What's performing well right now?"
"Are there any concerning trends?"
"Which buildings need attention?"