Answers stop before the work starts
A generic chatbot may explain what to do, but the project manager still has to open another system, copy the answer, and create the record manually.
Ask a question or request an outcome. The PlanSpan Agent searches reports, RFIs, submittals, drawings, specifications, schedules, photos, and files—then uses project tools to complete the next step.
WorkspaceCedar Grove Medical Centre
PlanSpan Agent
Read today's daily report
Issue found: missing firestopping at Level 3 riser
Searched project knowledge
Drawing A-601 · Specification 07 84 00
Drafted RFI
Question, context, references, and proposed response
Created RFI-027
Saved as draft for project-team review
RFI-027 · Firestopping at Level 3 riser
Draft created with references to A-601 and Section 07 84 00. Review before issuing to the consultant.
Why teams need an agent
A generic chatbot may explain what to do, but the project manager still has to open another system, copy the answer, and create the record manually.
The evidence behind one decision may sit in a daily report, drawing detail, specification clause, photo, previous RFI, and approved submittal.
Turning a field issue into a properly structured RFI, report, invoice, or punch item takes time and creates another opportunity to lose context.
Project-aware execution
PlanSpan combines project knowledge retrieval with application tools, so the same agent can investigate the request, perform the workflow, and return the completed record.
Search RFIs, submittals, daily reports, drawings, specifications, schedules, photos, invoices, and files from one request.
Distinguish between a question that needs an answer and an instruction that requires one or more project actions.
Read source records, collect supporting context, draft the content, and create the requested object inside PlanSpan.
Return source-aware results and direct links so the team can inspect the information used by the agent.
Prepare review-ready work while keeping project professionals responsible for what is approved or formally issued.
Use the signed-in user's project context and permissions when searching records or performing an action.
How it works
The agent plans the work, uses the right project tools, and returns a result your team can inspect.
Ask a project question or tell the agent what to create, update, or investigate in plain language.
The agent selects the relevant search tools and reads the project records needed to understand the request.
It invokes one or more application functions to create or update the requested project object.
The agent confirms what changed and returns a direct link to the new or updated record.
Related AI workflows
See how project documents become an active, searchable knowledge layer for the agent.
Draft an RFI response using the relevant drawings, specifications, and project history.
Compare a submittal with project requirements and prepare source-linked findings.
Frequently asked questions
A construction AI agent is software that can interpret a project request, search relevant construction records, decide which application tools to use, and carry out a workflow such as drafting an RFI or creating a daily report.
A chatbot primarily returns text. An agent can also call approved application functions, perform multiple steps, save the result to the project system, and link the user to the completed record.
The agent can work with project information available in PlanSpan, including RFIs, submittals, daily reports, drawings, specifications, schedules, photos, files, invoices, and other supported project records.
No. PlanSpan is designed to prepare source-aware, review-ready work while the project team remains responsible for checking, approving, and formally issuing documents.
Put project knowledge to work