The PlantDemand MCP server lets AI assistants answer real operational questions using your live scheduling data. This page collects prompts that plant managers, dispatchers, and operations leaders find most useful day-to-day. Each example shows the kind of question you can ask once your AI client is connected — see the quickstart for setup.
Daily dispatch questions
These are the questions dispatchers used to answer by clicking through the schedule interface. With MCP, they become natural language:
- “What is on the schedule for plant 14 tomorrow?”
- “List all orders scheduled for this week sorted by customer.”
- “Which jobs are scheduled before 7 AM next Monday?”
- “Show me every order date that has the ‘rush’ flag set this month.”
Capacity and production planning
Capacity questions used to require building a report. Now they are conversation:
- “What is our total scheduled tonnage at plant 14 for August?”
- “Compare scheduled tonnage by week for the next four weeks.”
- “Which day next week has the most scheduled production across all plants?”
- “What is the average daily scheduled tonnage at plant 14 over the past 30 days?”
Customer mix analysis
Understanding who you are running for becomes a one-line query:
- “Who are our top five customers by scheduled tonnage this month?”
- “How much DOT-flagged production is on the schedule this quarter?”
- “Which customers have orders scheduled at multiple plants?”
- “Show me all orders for customer ABC Construction in the next 30 days.”
Material and mix design queries
Mix design and material questions cut across the full schedule:
- “List every mix design scheduled at plant 14 this week.”
- “How many tons of Superpave 12.5 are scheduled this month?”
- “Which materials appear on more than 10 percent of orders this season?”
- “Show me orders that use a mix design we have not run in the past 90 days.”
Pattern and trend questions
This is where MCP earns its keep — questions that would take an analyst hours to answer become instant:
- “Which day of the week tends to have the most schedule changes?”
- “Compare this month’s scheduled tonnage to the same month last year.”
- “Identify any customers whose order volume has dropped by more than 20 percent quarter over quarter.”
- “Which mix designs have the longest production runs on average?”
Cross-functional workflows
The most powerful use cases combine MCP with other AI capabilities the assistant already has:
- “Pull this week’s scheduled tonnage from PlantDemand and summarize the key changes from last week in an email I can send to ownership.”
- “Look at next week’s schedule and flag any days where we have back-to-back specialty mixes that might require a clean-out between runs.”
- “Compare our scheduled production to the regional weather forecast and highlight any days with significant rain risk.”
Building good prompts for MCP
A few patterns make MCP-backed prompts more reliable:
- Be specific about plants and dates. “What’s on the schedule” is ambiguous. “What is on the schedule for plant 14 next Monday” is unambiguous and the agent will produce the right tool call.
- Use plant IDs when you know them. The agent can look up plants by name through the list-plants tool, but using the ID directly is faster and removes ambiguity when plant names are similar.
- Ask the agent to cite the data. Add “show me which records you used” to the prompt and the agent will reference the underlying tool responses.
- Pair with a system prompt. A short system prompt that tells the agent to call
initializefirst and to never invent values significantly improves accuracy. See the recommended template in the quickstart.
Where to go next
- Quickstart — Set up the connection
- Tool reference — See what data is available
- Client setup guides — Wire MCP into your assistant
- AI for asphalt plant scheduling — Background on what AI can and cannot do in plant operations
This guide is part of the PlantDemand hub for asphalt plant operations, scheduling, and sales management.