Construction data & AI, in plain language.

An AR aging report groups open customer invoices into buckets by days outstanding, typically current, 1-30, 31-60, 61-90 and over 90 days. In construction it usually tracks pay applications and retainage receivable by project and owner. It is the starting point for collections work and cash forecasting.

Why it matters: Slow-paying owners strain cash and bonding capacity long before they show up as a loss. Seeing aging by project lets controllers chase the right invoices before cash gets tight.

Procore + QuickBooks + HubSpot reporting →Solutions for controllers →

Backlog is the remaining value of signed contracts that has not yet been recognized as revenue, usually calculated as contract value (including approved change orders) minus revenue earned to date. Many contractors also track how quickly backlog burns down over time. It is a core input to capacity planning and to surety and lender conversations.

Why it matters: Backlog tells leadership how much work is locked in and when it will run out. Pairing it with burn rate shows whether the company needs to win work now or protect capacity.

Construction WIP reporting in Power BI →Solutions for operations leaders →

A change order modifies the original contract, adjusting scope, contract value, schedule or a combination. Changes usually move through stages such as potential change, change order request and approved change order. Approved owner change orders flow into contract value, while subcontract change orders flow into committed cost.

Why it matters: Unapproved or untracked changes are a common source of margin fade and billing disputes. Seeing pending change exposure by project protects both revenue and cash.

Procore API integration guide →Solutions for controllers →

Committed cost is the total value of subcontracts and purchase orders issued on a job, including approved commitment change orders. It is often split into amounts already invoiced and amounts still open. Committed cost plus uncommitted estimates feeds the forecast of total cost at completion.

Why it matters: Commitments show where money is already spoken for before invoices arrive. Ignoring them makes a job look healthier than it is.

Construction WIP reporting in Power BI →

Cost codes break a project budget and actual costs into categories of work, such as concrete, framing or electrical, often aligned to CSI MasterFormat. They are typically combined with cost types like labor, material, equipment and subcontract. Consistent cost codes are what let estimates, budgets, commitments and actuals line up.

Why it matters: When cost codes differ between the estimate, the PM system and accounting, cost reports stop reconciling. A shared code structure, or a crosswalk between them, is the foundation of reliable job cost reporting.

Construction data quality rules →Crosswalk →

The cost-to-cost method measures progress as job-to-date cost divided by the current estimated total cost at completion. That percentage is applied to the contract value to calculate revenue earned. It is a common input method for recognizing revenue over time on construction contracts.

Why it matters: Revenue, profit and over/under billing all depend on this number. If the estimate at completion is stale, earned revenue and margin are wrong too.

Construction WIP reporting in Power BI →WIP schedule →

Estimate at Completion is the projected final cost of a project, usually cost to date plus the estimated cost to complete the remaining work. Project managers update it as quantities, productivity, commitments and changes become known. It drives percent complete, projected margin and over/under billing.

Why it matters: A stale EAC hides problems until late in the job, when options are limited. Regular, evidence-based EAC updates are what make a WIP schedule trustworthy.

Construction WIP reporting in Power BI →Fade and gain →

Fade is a decrease in projected gross profit or margin on a job over time; gain is an increase. It is measured by comparing the current projected margin to the margin at bid or at a prior period. Fade often comes from cost overruns, unapproved changes or optimistic early estimates.

Why it matters: Consistent fade across jobs points to estimating or project control problems, and sureties watch it closely. Catching it early gives the team time to recover margin.

Construction WIP reporting in Power BI →Solutions for controllers →

Job costing records labor, material, equipment, subcontract and other costs against individual projects, usually by cost code and cost type. A job cost report compares budget, committed cost, actual cost and forecast for each line. It is the core financial view of how a project is performing.

Why it matters: Accurate job cost is what lets a contractor know which jobs make money and why. Late or miscoded costs distort every downstream report.

Construction WIP reporting in Power BI →Cost code →

Overbilling (billings in excess of costs and estimated earnings) means a contractor has billed more than it has earned, which is recorded as a liability. Underbilling (costs and estimated earnings in excess of billings) means it has earned more than it billed, recorded as an asset. Both are calculated per job on the WIP schedule from percent complete, contract value and billings to date.

Why it matters: Large underbillings can signal unbilled work, unapproved changes or cost overruns, and sureties scrutinize them. Overbillings can mask cash that will be needed to finish the job.

Construction WIP reporting in Power BI →WIP schedule →Solutions for controllers →

A pay application is a progress billing submitted to the owner, commonly monthly, showing work completed and materials stored against each line of the schedule of values. The AIA G702 and G703 forms are a widely used format. It also shows retainage withheld and previous payments.

Why it matters: Pay app accuracy and timing directly drive cash flow. Rejected or late pay apps push collections back a full cycle.

Procore + QuickBooks + HubSpot reporting →Retainage →

Retainage is a percentage of each progress payment that the owner holds back, and that the contractor in turn holds from subcontractors, as security for completing the work. The rate is set by contract and sometimes limited by state law, and is often in the range of 5 to 10 percent. It is released at substantial or final completion, sometimes reduced partway through.

Why it matters: Retainage can tie up a large share of a contractor's cash across the portfolio. Tracking retainage receivable and payable by job helps get it released on time.

Construction project reporting showcase →Pay application →

A schedule of values divides the contract sum into line items, often by trade or work area, each with a scheduled value. Progress is billed against each line on the pay application. Approved change orders are added as new or adjusted lines.

Why it matters: A well-structured schedule of values makes billing faster and disputes rarer. It also lets billing progress be compared with cost and schedule progress.

Pay application →Solutions for controllers →

A work-in-progress schedule lists each open contract with contract value, estimated total cost, cost to date, percent complete, earned revenue, billings and the resulting over or under billing. It also shows projected gross profit and often backlog. It is prepared at least quarterly, often monthly, and is reviewed by sureties, lenders and CPAs.

Why it matters: The WIP schedule is one of the most closely watched views of contractor financial health. Building it from source systems turns a manual close task into a repeatable report.

Construction WIP reporting in Power BI →Procore + QuickBooks + HubSpot reporting →Solutions for controllers →

A baseline is a saved copy of the approved project schedule, including planned start and finish dates for every activity. Updates are compared to the baseline to measure variance and slippage. In Primavera P6 a project can store multiple baselines, and contracts often require a specific one for delay analysis.

Why it matters: Without a clean baseline, there is no defensible way to show what slipped, when and why. That matters for owner reporting and for any time-extension claim.

AI for CPM scheduling and P6 →Scheduling topic →

The Critical Path Method models a project as activities with durations linked by logic relationships, then calculates early and late dates for each one. The longest chain of dependent activities sets the earliest possible finish date. Tools like Primavera P6 run these calculations when the schedule is scheduled.

Why it matters: CPM tells the team which activities can slip without moving the finish date and which cannot. It is also the basis most contracts use to evaluate delays.

AI for CPM scheduling and P6 →Connect agentic scheduling demo →

The critical path is the sequence of logically linked activities with the least total float, usually zero, running from project start to finish. Any delay to a critical activity delays project completion unless the work is resequenced or accelerated. In P6 it can be defined by total float or by the longest path.

Why it matters: Knowing which activities are critical tells superintendents where a day lost is a day lost on the project. It focuses recovery effort and supports delay claims.

AI for CPM scheduling and P6 →Primavera P6 MCP server →Solutions for scheduling teams →

A Definable Feature of Work is a task that is separate and distinct from other tasks and has its own control requirements, a concept used in USACE and NAVFAC three-phase quality control. Each DFOW goes through preparatory, initial and follow-up inspection phases. DFOWs are often tied to schedule activities so quality and schedule progress can be tracked together.

Why it matters: Missing a preparatory or initial phase inspection can stop work or trigger rework on federal jobs. Tracking DFOWs alongside the schedule shows which features are ready to start.

Fabric construction reporting demo →Building construction reporting in Microsoft Fabric →

A driving path traces backward from a chosen activity or milestone through the predecessors that actually determine its start or finish date. It is like a critical path aimed at an interim target instead of the project end. P6 can show driving relationships and trace logic to find it.

Why it matters: Interim milestones such as dry-in or turnover often carry their own commitments. The driving path shows exactly which work has to move to protect them.

AI for CPM scheduling and P6 →Critical path →

Total float is the amount of time an activity can be delayed without delaying the project finish or a constrained date. Free float is how long it can slip without delaying any immediate successor. Activities with zero or negative total float are critical or already behind.

Why it matters: Float shows where there is room to absorb problems and where there is none. Watching float erode over updates is an early warning of schedule trouble.

AI for CPM scheduling and P6 →Critical path →

An XER file is a tab-delimited text export from Primavera P6 containing tables for projects, WBS, activities, relationships, calendars, resources and more. It is the most common way schedules are shared between contractors, owners and consultants. Because it is plain text, it can be parsed and analyzed without a P6 installation.

Why it matters: Many schedule submissions arrive as XER files. Parsing them directly enables automated schedule quality checks and analysis without manual re-entry.

Primavera P6 MCP server →P6 MCP server demo →

An API is a set of endpoints and rules that lets one system request or send data to another, usually over HTTPS with JSON payloads and authentication such as OAuth. Construction platforms like Procore, QuickBooks Online and HubSpot expose REST APIs for projects, costs, contacts and more. Integrations, data pipelines and AI tools are all built on top of APIs.

Why it matters: If a system has a usable API, its data can flow into reporting and automation instead of being retyped. API limits and permissions shape what an integration can realistically do.

Procore API integration guide →Integrations topic →

Bronze holds data as it arrived from source systems, kept for traceability and reprocessing. Silver holds cleaned, typed, de-duplicated and conformed data, with keys mapped across systems. Gold holds business-ready tables such as facts and dimensions shaped for reporting and AI.

Why it matters: Separating layers means a bad source record can be traced and fixed instead of silently corrupting a report. It also lets new reports reuse the same trusted gold tables.

Building construction reporting in Microsoft Fabric →Microsoft Fabric for contractors →

A crosswalk maps identifiers between systems, for example a Procore project ID to a Sage job number, or a PM cost code to an accounting cost code. It is maintained as an explicit table rather than buried in report logic. Records that fail to match are flagged for review instead of being dropped.

Why it matters: Most cross-system reports break on mismatched keys, not on math. A maintained crosswalk is what lets project, cost and schedule data roll up to one version of each job.

Construction data quality rules →Building construction reporting in Microsoft Fabric →

A data-quality gate is a set of rules run during a pipeline, such as required fields present, keys matched, values in range, or totals reconciling to source. Records that fail are flagged with a reason and surfaced to an owner, rather than silently dropped or passed through. Gates can block a refresh when failures exceed a threshold.

Why it matters: One unmapped job or duplicated invoice can make an executive report wrong. Gates make problems visible and fixable at the source.

Construction data quality rules →Our approach →

Direct Lake is a semantic model storage mode in Microsoft Fabric that loads data on demand from Delta tables in OneLake, without a scheduled import copy. It aims to combine near-import query performance with data that reflects the latest lakehouse tables. Some conditions cause queries to fall back to DirectQuery or fail, so models need to be designed for it.

Why it matters: Reports can reflect new pipeline data without waiting on a separate dataset refresh. That shortens the gap between a source update and what leadership sees.

Microsoft Fabric for contractors →Semantic model →

ETL extracts data from source systems, transforms it by cleaning, mapping and calculating, and loads it into a destination such as a warehouse or lakehouse. Many modern platforms use ELT, landing raw data first and transforming it inside the platform. In Microsoft Fabric this is typically done with pipelines, notebooks and dataflows.

Why it matters: Well-built pipelines replace manual exports and spreadsheet merges that eat days each month. They also make every number traceable back to its source.

Building construction reporting in Microsoft Fabric →Reporting topic →

A lakehouse combines the low-cost file storage of a data lake with table features of a warehouse, such as schemas and transactions, typically using open formats like Delta Lake. In Microsoft Fabric a lakehouse stores data in OneLake and exposes a SQL analytics endpoint. It can feed reports, notebooks and AI from the same copy of data.

Why it matters: Contractors can bring ERP, project management and schedule data into one governed place instead of a pile of exports. Every report then starts from the same numbers.

Microsoft Fabric for contractors →Medallion architecture →

Medallion architecture organizes a lakehouse into progressive layers: bronze for raw source data, silver for cleaned and conformed data, and gold for business-ready models. Each step adds validation and structure. Because raw data is retained, any layer can be rebuilt when rules change.

Why it matters: It gives contractors a traceable path from a Procore or ERP record to a number on a report. Problems can be fixed once at the right layer instead of patched in every report.

Microsoft Fabric for contractors →Building construction reporting in Microsoft Fabric →Fabric construction reporting demo →

A semantic model, formerly called a Power BI dataset, holds the tables, relationships, calculations (DAX measures) and security rules that reports use. It gives business-friendly names and one shared definition for metrics like earned revenue or percent complete. Many reports can be built on the same model.

Why it matters: When metrics are defined once in a semantic model, every report and AI query uses the same math. That ends arguments about whose spreadsheet is right.

Microsoft Fabric for contractors →Direct Lake →Star schema →

A snapshot captures the state of records at a specific moment, for example the WIP schedule at each month-end close or schedule float at each update. Source systems usually only show the current state, so snapshots preserve history. They enable period-over-period comparisons and trend analysis.

Why it matters: Questions like how much margin faded since last quarter can only be answered if prior states were saved. Snapshots also make closed-period reports reproducible.

Construction WIP reporting in Power BI →Replace the monthly Excel report →

A star schema organizes data into fact tables holding measurable events, such as cost transactions or billings, and dimension tables describing them, such as project, cost code, vendor and date. Facts connect to dimensions through keys. It is the recommended modeling pattern for Power BI.

Why it matters: Star schemas make reports faster and measures simpler and more reliable. They also make it easy to slice the same numbers by project, PM, region or period.

Semantic model →Reporting topic →

Agentic AI describes systems where a language model decides which steps to take, calls tools or APIs, inspects the results and continues until a goal is met. Unlike a chatbot that only produces text, an agentic system can read from and write to real business systems. That power is why permissions, approvals and logging matter so much.

Why it matters: Agentic workflows can take repetitive coordination work off project teams, such as pulling schedule data or drafting updates. Without governance they can also make changes nobody reviewed.

Connect: AI agents for CPM scheduling →Governing AI agents in construction →

An AI agent combines a language model with a defined role, instructions, and a set of tools it is allowed to call, such as reading Procore RFIs or querying a P6 schedule. The agent decides which tools to use for a request and returns an answer or action based on the results. Good agents are narrow, well-scoped and auditable.

Why it matters: A focused agent can answer project questions from live system data instead of someone digging through five applications. Scope and permissions decide whether it is useful or risky.

Construction AI agents with MCP tools →Construct.Chat agents demo →Agents topic →

A golden set is a curated collection of representative inputs paired with expected outputs or acceptance criteria. Each change to prompts, tools or models is run against it to see whether quality improved or regressed. It turns subjective impressions of an AI system into repeatable measurements.

Why it matters: Without a golden set, nobody can say whether an agent got better or worse after a change. It is the evidence that an AI tool is ready for production.

Governing AI agents in construction →Rubric →

Human-in-the-loop means an AI system pauses for a person to confirm, edit or reject an action, especially writes to business systems. Read-only questions may run freely while changes such as updating a schedule or sending a message require approval. The approval and its outcome are logged.

Why it matters: It lets teams get the speed of AI while keeping accountability with the people who own the work. It is often the difference between a pilot and a production rollout.

Governing AI agents in construction →Governance →

Least privilege is a security principle that limits each identity to the minimum permissions required for its job. For AI agents that means exposing only the tools, projects and operations relevant to a role, with write access granted narrowly. Scoped API credentials and tool allowlists are common ways to enforce it.

Why it matters: If an agent is misled or makes a mistake, least privilege limits the damage. It is usually the first thing IT and security teams ask about.

Governance →Solutions for IT and security →

The Model Context Protocol is an open protocol that defines how AI applications discover and call tools, read resources and use prompts provided by external servers. An MCP server wraps a system such as Procore or P6, and any MCP-compatible client can use it. It replaces one-off integrations for each AI app with a shared interface.

Why it matters: MCP lets the same construction system connectors be reused across AI assistants and agents. That makes AI work against live project data practical and governable.

What is MCP for construction →Procore MCP architecture demo →

An MCP gateway sits in front of one or more MCP servers and handles concerns like authentication, permission checks, tool routing, rate limiting and telemetry. Clients connect to the gateway rather than to each server directly. This centralizes policy and gives a single record of what tools were called.

Why it matters: As tool counts grow, a gateway keeps access consistent and auditable across agents. IT teams get one place to enforce and review policy.

MCP gateway telemetry and tool runtime →Tool Runtime gateway demo →

An MCP server implements the Model Context Protocol for a specific system, describing its available tools with names, inputs and descriptions. When an AI client calls a tool, the server performs the underlying API call or query and returns the result. Servers can run locally or remotely.

Why it matters: A well-built MCP server lets an AI assistant work with a scheduling or project system directly instead of relying on copied data. Tool descriptions and permissions determine how reliable it is.

Primavera P6 MCP server →P6 MCP server demo →

Retrieval-augmented generation finds relevant passages from a document collection, often using vector embeddings, and gives them to a language model along with the question. The model answers from that retrieved context rather than only from its training. Good RAG systems cite the sources they used.

Why it matters: RAG lets teams ask questions of specs, contracts and procedures with answers tied to the actual document. Citations make it possible to verify the answer.

Construction AI agents with MCP tools →Vector embeddings →

A rubric defines what a good answer looks like, such as correct figures, cited sources, the right tool used and no unsupported claims, often with a score for each criterion. It can be applied by people or by an automated grader. Rubrics are paired with a golden set to evaluate AI systems consistently.

Why it matters: A rubric turns opinions about an agent into criteria the business agrees on. It makes go-live decisions and regression checks objective.

Governing AI agents in construction →Golden set →

Telemetry is the automatic collection of operational data from a running system. For AI agents and MCP tools it typically includes which tools were called, by whom, with what inputs, how long they took and whether they succeeded. It feeds monitoring, auditing and cost tracking.

Why it matters: Telemetry is how a team proves what an agent did and spots failures before users complain. Security and compliance reviews depend on it.

MCP gateway telemetry and tool runtime →Governance →

Tool routing narrows a large set of available tools down to the ones relevant to a specific request before the AI model chooses among them. It can use categories, search, embeddings or rules. Presenting fewer, better-matched tools improves accuracy and reduces cost.

Why it matters: Large platforms like Procore expose thousands of API operations, far too many to hand an AI model at once. Routing makes large tool catalogs usable.

Routing 2,755 Procore tools →Procore MCP architecture demo →

An embedding model converts text into a list of numbers, a vector, so that passages with similar meaning end up close together. Stored in a vector index, embeddings support semantic search that matches concepts rather than exact keywords. They are a core building block of RAG and of some tool routing approaches.

Why it matters: Embeddings let someone ask about waterproofing at the podium and find the right spec section even if the wording differs. That makes large document sets searchable in plain language.

RAG →Tool routing →

HubSpot is a customer relationship management platform with tools for marketing, sales and service. Contractors use it to track prospects, bids and opportunities through a sales pipeline. Its API lets deal and pipeline data be combined with project and financial data.

Why it matters: Joining pipeline data with backlog and capacity shows whether upcoming wins fit the crews and cash available. That turns sales forecasting into operational planning.

Procore + QuickBooks + HubSpot Fabric build →Procore + QuickBooks + HubSpot demo →

Microsoft Fabric is a SaaS analytics platform that brings together data integration, lakehouses, warehouses, notebooks, real-time analytics and Power BI on shared OneLake storage. Workloads share security and capacity. It is a common choice for organizations already using Microsoft 365 and Power BI.

Why it matters: Fabric lets a contractor build a governed data platform without stitching together many separate services. Reports, pipelines and AI can share one copy of the data.

Microsoft Fabric for contractors →Fabric construction reporting demo →

n8n is a workflow automation tool where users build flows from nodes that trigger on events, call APIs and transform data. It can be self-hosted or used as a cloud service, and it supports custom nodes. Teams use it to automate handoffs between business systems.

Why it matters: Custom construction nodes let operations teams automate routine handoffs, such as syncing records or sending alerts, without a full software project each time.

n8n construction nodes demo →Integrations topic →

Outbuild is a cloud scheduling platform for construction teams that supports master schedules and collaborative short-term planning. It is used to coordinate look-ahead and weekly work plans with field teams. Its data can be combined with cost and project management data for reporting.

Why it matters: Bringing field planning data next to cost and project data shows whether planned work is actually getting done. That connects schedule commitments to financial outcomes.

Building construction reporting in Microsoft Fabric →Scheduling topic →

Power BI is Microsoft's analytics and reporting service, made up of semantic models, reports and dashboards shared through the Power BI service. It uses DAX for calculations and supports row-level security. It is also one of the workloads inside Microsoft Fabric.

Why it matters: Most contractors already have Power BI through Microsoft 365, so well-modeled reports can reach executives and PMs without new software. The model underneath decides whether the numbers can be trusted.

Monthly construction report: Procore to Power BI →Construction project reporting showcase →

Oracle Primavera P6 Enterprise Project Portfolio Management is scheduling software for planning and controlling large projects using CPM. It manages WBS, activities, relationships, calendars, resources and baselines across portfolios. Schedules are commonly exchanged as XER files.

Why it matters: Many owners and contracts require P6 schedules, so schedule data lives there. Connecting it to cost and field data turns the schedule into a live management tool.

Primavera P6 MCP server →Procore + P6 Power BI demo →

Procore is a construction management platform covering project management (RFIs, submittals, drawings), quality and safety, and project financials such as budgets, commitments and change orders. It exposes a REST API documented at developers.procore.com. Many contractors use it as the system of record for field and project data.

Why it matters: Procore holds much of the day-to-day project record. Getting that data into reporting and AI tools, reliably and with the right permissions, unlocks most cross-system insight.

Procore API integration guide →Routing 2,755 Procore tools →

QuickBooks Online is cloud accounting software for general ledger, invoicing, bills and payments. Contractors often use projects, classes or customers to track job costs. Its API allows financial data to be combined with project management and CRM data.

Why it matters: For growing contractors, QuickBooks often holds the actual cost and billing data that project systems lack. Joining the two is the basis of WIP and cash reporting.

Procore + QuickBooks + HubSpot reporting →Procore + QuickBooks + HubSpot demo →

Sage 100 Contractor is a construction accounting system, typically run on-premises, covering general ledger, job cost, payroll, AP and AR, and service. It stores jobs, cost codes and cost types used for job cost reporting. Its data is often extracted into a data platform to combine with project management systems.

Why it matters: Accounting holds the actual costs and billings, so WIP and margin reporting depend on it. Mapping its jobs and cost codes to the project system is a key integration step.

Building construction reporting in Microsoft Fabric →Crosswalk →