What You Can Ask Mosaic
Example prompts, the full list of Mosaic data and reports your AI assistant can pull, and how to phrase questions for accurate answers.
Once your assistant is connected, you can ask about anything in your Mosaic workspace — projects and phases, who's assigned where, logged hours, capacity, budgets, and rates — without opening Mosaic or building a report.
Haven't connected yet? Start with Connect Your AI Assistant to Mosaic.
How Your Assistant Answers
Your assistant doesn't search text — it runs structured lookups and reports against Mosaic, the same data that powers the Mosaic app. For each question it decides:
- Which lookup or report answers it — projects, members, time entries, utilization, budget tracking, and so on.
- What to filter by — date range, project, phase, client, member, role, or billable status.
- How to group the results — by person, project, phase, activity, role, client, or team, or a combination like member per project or role per phase.
- What time breakdown to use — a single total, or a day-by-day, week-by-week, month-by-month, or year-by-year split.
- Which unit to report in — hours, full-time equivalents (FTEs), bill rates, or cost rates.
You don't need to know any of that to ask a question. But naming the pieces you care about — the period, the grouping, the unit — gets you a more precise answer on the first try.
Example Prompts
Projects, Phases, and Portfolios
- 💬 "List all active projects for Acme."
- 💬 "What phases does the Riverside Tower project have, and what are their start and end dates?"
- 💬 "Which projects are in the Q3 Delivery portfolio?"
People, Roles, and Skills
- 💬 "Who's on the Riverside Tower project, and what role does each person have?"
- 💬 "Which team members have Revit skills?"
- 💬 "List everyone in the Chicago office with a Senior Engineer role."
Availability, Capacity, and Scheduling
- 💬 "Who has capacity next week for 20 hours of design work?"
- 💬 "Show me who's over capacity in September."
- 💬 "What's on Priya's schedule for the next two weeks?"
Time and Timesheets
- 💬 "How many hours did the team log to Acme last month?"
- 💬 "Show me billable versus non-billable hours by person for July."
- 💬 "Whose timesheets are still unsubmitted for last week?"
Budgets, Rates, and Profitability
- 💬 "Which projects are over budget this quarter?"
- 💬 "What bill rate is each person on Riverside Tower charged at?"
- 💬 "What's the profit margin on our Acme projects year to date?"
Multi-Step Analysis
Your assistant can chain several lookups together and reason across the results.
- 💬 "Compare planned versus logged hours by phase on Riverside Tower for Q3, and flag anything more than 20% over."
- 💬 "Show utilization by role for the last six months, month by month, and tell me which roles trended down."
- 💬 "Cross-check the projects that are over budget against the people scheduled on them next month."
Reports Your Assistant Can Pull
Mosaic exposes thirteen built-in reports. Each one accepts a date range, filters (projects, phases, clients, members, roles, billable status), and a grouping level.
| Report | Answers questions like | Notes |
|---|---|---|
| Utilization | How busy is the team? What share of available capacity did we use? | Spent hours — total, billable, non-billable, and PTO — against capacity. Supports time-interval breakdowns. |
| Workload | Who's overloaded, and who has room? | Planned hours split by assigned/unassigned and tentative/confirmed, against capacity. Supports interval breakdowns. |
| Role Demand | Which roles do we need more of? | Role-level demand with FTE conversions. Reports in hours or FTEs. |
| Spent Time | How many hours were actually logged? | Logged hours and expense at bill or cost rates, with no capacity comparison. |
| Planned Time | How much work is scheduled? | Planned hours from work plans — future scheduled work, not logged hours. |
| Total Time | How do estimated, spent, and planned hours compare? | All three time types in one view. |
| Budget Tracking | Where does the budget stand right now? | Budget fee and hours, estimated, spent, planned, remaining, and profit. Supports interval breakdowns; reports in hours, FTEs, bill rates, or cost rates. |
| Budget Variance | How far off budget are we? | Budgeted amounts against actual spent and planned. Single-period totals. |
| Profit Margin | How profitable is this project? | Fee, profit, remaining budget, and margin percentages. Single-period totals, at bill or cost rates. |
| Backlog | What work is still outstanding? | Planned work not yet completed, with remaining budget. Single-period totals. |
| Time Variance | Are we over our estimates or our plan? | Estimated, spent, and planned hours with spent-vs-planned and spent-vs-estimated variances. Supports interval breakdowns. |
| Schedule Variance | Are we behind schedule? | Time variance plus baseline start and end dates and workday calculations. Baselines available for projects, phases, and roles within a phase. |
| Timesheet Status | Whose timesheets are outstanding? | Hours by submission status — not submitted, submitted, rejected, approved — with a billable split. Can be grouped by member or team. |
Tip: Reports always need a date range. Saying "Q3 2026" or "July 1 to September 30" gets you a cleaner answer than "recently."
Data Your Assistant Can Look Up
Beyond reports, your assistant can pull more than sixty individual record lookups. Here's what's available by area:
| Area | Includes |
|---|---|
| Projects & portfolios | Projects, phases and sub-phases, work categories within phases, portfolios, clients, project scopes, and dependencies |
| Tasks | Project tasks, task lists, and task assignments |
| Project teams | Who's assigned to each project, their project-level roles, and their project rate breakdowns |
| People | Team members, their team-level and secondary roles, and their skills |
| Organization tags | Departments, offices, regions, and disciplines — including which members and which projects carry each tag |
| Skills | The team skill catalog, member skills, and the skills a role or assignment requires |
| Scheduling & capacity | Work plans (who's scheduled on what, and when), capacity policies, holiday policies, time-off policies, and daily check-ins |
| Time | Time entries with logged hours and billable status |
| Rates & budgets | Bill rates, cost rates, member and role rate assignments, the rate catalog, project rate overrides, currency exchange rates, budget estimates, and per-phase fee breakdowns |
| Activity history | An audit trail of who changed what and when across projects, phases, tasks, work plans, time entries, and memberships |
Your permissions apply. Every request is checked against your Mosaic permissions, so your assistant can only return data you can already see in the Mosaic app.
Getting Better Answers
Be specific about the period. Reports need a date range. "Q3 2026" or "the last 8 weeks" works; "lately" forces the assistant to guess.
Name the grouping you want. "By person", "by phase", "by role and month" — otherwise you may get a single total when you wanted a breakdown.
Say which unit. "In hours", "in FTEs", "at bill rates", "at cost rates". This matters most for budget and capacity questions.
Filter in the question, not after. "Acme projects only" is faster and more accurate than pulling every project and narrowing afterwards.
Break big questions into steps. Ask for the list first, then the analysis. A single sprawling request is more likely to time out or return partial data.
Ask it to show its work. "Which projects did you include?" or "What date range did you use?" makes it easy to catch a wrong assumption.
Say what format you want. A table, a bulleted summary, or CSV you can paste into a spreadsheet.
Spot-check before you share. Verify any number against Mosaic before it goes into a client update or a board deck.
Things to Know
- Large questions can be slow or partial. Results are paginated, so very broad queries may need several passes. If a list looks short, ask "did you get all of them?" — or narrow the question.
- Answers reflect Mosaic at the moment you ask. Ask again for fresh numbers; your assistant doesn't watch for changes.
- Results depend on your Mosaic data. Utilization needs logged time; budget reports need budgets entered. Gaps in Mosaic show up as gaps in the answer.
- Assistants can misread or over-summarize. They're good at pulling and framing data, and less reliable at arithmetic across many records. Verify anything that drives a decision.
- Quality varies by model. More capable models handle multi-step questions and large result sets noticeably better.
Updated about 24 hours ago