The Sitemate MCP server connects AI tools directly to your workspace data — including projects, teams, templates, and form submissions.
This MCP prompt guide shows you how to write clear, specific queries so AI tools return accurate and complete results from your Dashpivot data.
Get Started in 3 Steps
You don't need to learn anything technical. Just follow these three steps:
Step 1 — Find out what's available
"What workspaces do I have access to? List the projects and teams."
Then:
"What form templates are available in [team name]?"
Step 2 — Ask for specific forms using the exact template name
"Show me one Daily Timesheet (w Tracking) from Team A, Project 1. Include all fields and values."
Step 3 — Narrow it down by date
"List all Daily Timesheet (w Tracking) forms from Team A, Project 1 created between 2026-02-01 and 2026-02-14. For each one, show the form ID, date, and employee name."
That's it! You're up and running.
Copy-Paste prompt Template
Use this structure:
I need to [what you're looking for] for [project or team name].
Template: [exact template name, e.g. "Daily Timesheet (w Tracking)"]
Time period: [e.g. this week / last month / 2026-02-01 to 2026-02-14]
Show me: [e.g. top 10 results / a summary / specific fields like employee name and hours]
Let me know if the results are complete or if some might be missing.
Example:
Show me the top 10 Daily Timesheet (w Tracking) forms from Team A, Project 1 for the first two weeks of February 2026, with employee name and hours worked. Let me know if the results are complete or if some might be missing.
Three important rules and prompting best practices
1. Always use the exact template name
Be specific — include template name, project/team, and date range.
AI tools query one template at a time — not all templates in a team. If you say “show me all forms,” you’ll only get results from one or two templates.
Instead of:
“Show me all forms from Team A”
Use:
“Show me all Daily Timesheet (w Tracking) forms from Team A”
Need data from multiple templates?
Run separate queries for each one.
Not sure what templates exist?
Ask: “What templates are available in [team name]?”
2. Keep your date ranges short (2 weeks or less)
If you need a longer period, just break it into 2-week chunks and ask multiple times.
1–7 days: very reliable
7–14 days: reliable
14–30 days: may miss some forms
30+ days: often incomplete
If you need longer periods, break them into smaller date ranges and query separately.
3. Double-check any totals or calculated numbers
Totals, overtime, and other formula-based fields may not always be accurate. Raw values are reliable—calculate totals if needed.
Key limitations to be aware of
Date range limits – Shorter ranges (2 weeks or less) are most reliable; results may be incomplete beyond 30 days
Auto-calculated fields – Totals and formula-based values may be inaccurate; use raw data where possible
Result limits per request – Not all results appear in one response; ask “Show me the next 50 results”
Text-only data access – Photos, signatures, sketches, and attachments are not supported
AI accuracy – Responses may not always be 100% accurate; verify important data
More Ways to Ask
Here are some ready-to-use question templates for common tasks:
See what's available:
"What workspaces do I have access to? List the projects and teams.""What form templates are available in [team name]?"
Pull up a specific form:
"Show me one [template name] from [team], [project]. Include all fields and values."
Get a list of forms from a date range:
"List all [template name] forms from [team], [project] created between [start date] and [end date]. For each one, show the form ID, date, and [the fields you care about]."
Compare across teams:
"Compare [template name] submissions across Team A, Team B, and Team C in [project]. How many forms did each team submit between [start date] and [end date]?"
Get a top-N summary:
"Show me the top 10 [template name] forms from [team/project] for [time period], sorted by date. Include [specific fields] and tell me if the results are complete or if some might be missing."
Investigate something specific:
"I noticed [something unusual] in [template name] forms from [team]. Show me examples from [date range]. Focus on [the field that seems off]. Are these typical of the full dataset?"
What to do when something goes wrong
No results found
Check the exact template name. Try asking "What templates are available in [team]?" first.
Check that the date range actually has data — try a broader range.
Double-check the spelling of project and team names.
Remove extra filters and start with a simple question first.
The results seem incomplete
Ask: "What template did you query?" — it may have only checked one.
Try a shorter date range (2 weeks or less).
Ask for the next page: "Show me the next 50 results."
Narrow the scope: "Just Team A instead of all teams."
The totals don't look right
Totals, overtime sums, and formula-based fields aren't always accurate. Trust the raw values you entered and calculate totals yourself.
The results don't match what I expect
Ask: "What date field and filter did you use?"
Ask: "Did you check all relevant templates or just one?"
Try sorting differently: "Sort by date instead."
Spot-check a form: "Show me all the field values for form [ID]."
Review a single form to verify data
Common Mistakes and Easy Fixes
Asking for “all forms” without specifying a template
Using broad date ranges (30+ days)
Asking multiple unrelated questions at once
Relying on calculated totals
Using vague terms like “recent” instead of specific dates
Frequently Asked Questions
How do I find out what templates exist in my team?
Ask: "What form templates are available in [team name], [project name]?" —
Why did I only get a couple of forms when I know there are way more?
AI tools query one template at a time. If you didn't specify a template name, one was selected automatically. Always include the exact template name.
How many forms can I get at once?
Up to 50 per page. Say "Show me the next 50" to see more. Remember — each question only covers one template type.
Can AI tools look at photos, signatures, or attachments?
No — only text-based fields can be read. Photos, sketches, signatures, and file attachments are not supported.
How accurate are auto-calculated fields (totals, overtime, etc.)?
They can be inconsistent. Values entered manually are reliable, but calculated fields should be double-checked.
What's the best date range to use?
Two weeks or less gives you the most reliable results. Going beyond 30 days often means some forms get missed.
Can I compare data across different projects?
Yes! Just make sure you specify the exact project and team for each comparison. Keep the scope focused for best results.
What does "Calling tool" mean?
It means the AI tool is fetching data from Dashpivot. You may need to approve the request (e.g. clicking "GetForms").
Advanced: Using Path References
If you're comfortable with Dashpivot's internal paths (they look like /ws_xxx/p_xxx/...), you can use them to be extra precise about which workspace, project, or team you're querying. This is entirely optional — everything above works without paths.
Help me [task] for [scope].
Scope:
Path: /ws_xxx/p_xxx/...
Template: [exact template name]
Date range: [start] to [end] (keep to 14 days or less)
Return:
[summary / table / top N]
Key field values as evidence
Whether the results are complete or partial
If the question is too broad, suggest how to narrow it