This assistant answers questions about space utilization, occupancy, and attendance using your organization's InnerSpace data. It works best when your question specifies what you actually want to know — a metric (attendance, occupancy, peak load, etc.), a timeframe, and the locations involved — rather than a vague term like "usage." Below are guidelines for phrasing questions to get accurate, useful answers.
1. Be explicit about intent when asking about "utilization" or "usage"
The assistant guesses which metric you mean based on context. Guide it directly:
You care about | Say something like |
Cost/ROI, is space worth the spend | "time in use" or "% of space in use" |
Attendance/visitors | "unique people" or "headcount" |
Crowding/comfort | "occupancy %" |
Worst-case/peak load | "peak occupancy" or "average of daily peaks" |
Supply vs demand for a space type | "% of rooms in use" |
If you just say "utilization," you'll get a best guess with a one-line note on how it was interpreted — check that note and correct if needed.
2. Be specific about timeframe
State the date range or period you want (e.g., "last quarter," "June to August," "Tuesdays over the past month"). If you don't specify, the assistant will pick a default window, which may not match what you had in mind. For time-of-day questions (e.g., "9am–12pm"), local timezone and business-hours filtering are handled automatically — you only need to say more if you want a different window (e.g., "include evenings" or "include weekends").
3. Be specific about location
Name the building, floor, or space type you want included — or excluded. Examples:
"Just the 3rd floor, not the whole building"
"Meeting rooms only, exclude desks"
"Compare Building A and Building B"
If you don't specify, results may span more (or less) than you intended, especially across multi-floor or multi-building portfolios.
4. For space/area questions
Default scope is workspace only (desks, offices, meeting rooms, collaboration areas). If you want support spaces or circulation area included, say so explicitly.
5. For forecasts and what-if scenarios
State clearly whether you're asking:
"What will happen if nothing changes" → forecast
"What would happen if X changed" → simulation (e.g., "+1 day/week onsite," "+30 headcount," "convert 2 focus rooms to a meeting room")
Good simulation prompts specify:
The lever that's changing (headcount, days onsite, room count/type)
Any threshold you care about (e.g., "flag if any room type exceeds 80% peak use")
The space types and locations you care about most, if you want the analysis focused
Example: "If we add 40 people three days a week on Floor 3, when do meeting rooms become a bottleneck, and what's the headroom before that happens?"
You'll get: a decision-focused answer, scenario numbers, the first constraint and how much room you have, and a confidence label — with technical detail (model choice, error metrics) at the bottom, not buried in the main answer.
6. For reports and charts
Ask directly — "turn this into a report" or "chart this" — after getting a data answer, and it'll generate a PDF or visualization from the same results rather than re-pulling data.
7. If something looks off
If a number seems inconsistent (e.g., occupancy % vs raw counts don't line up), say so directly — the assistant is built to flag and explain anomalies rather than paper over them, and pointing it out gets you a faster answer than rephrasing the original question.