AI agents in Dynamics 365 Project Operations: Time, expense, and approvals

Project teams spend a surprising amount of time on work that does not move a project forward. Entering hours, processing receipts, checking submissions, and chasing missing information can take time away from customers, project delivery, and financial oversight. AI agents in Dynamics 365 Project Operations are designed to take on parts of this repetitive work.

Microsoft currently provides agents for time entry, expense processing, and approvals. Copilot Studio gives organizations another path for building custom agents around their own processes. The opportunity is not to automate every decision. It is to identify routine work that AI can complete or prepare. This allows employees and project managers to focus their attention where judgment is needed.

What is Dynamics 365 Project Operations?

Microsoft Dynamics 365 Project Operations helps project-based organizations manage work from the initial opportunity through project planning, resourcing, delivery, time and expense capture, billing, and financial management. Instead of maintaining separate systems for sales, project managers, consultants, resource managers, and finance teams, Project Operations can connect these activities around the same project information. This gives organizations better visibility into project schedules, resource use, costs, revenue, and profitability.

What are AI agents in Dynamics 365 Project Operations?

AI agents are software-based assistants that can perform tasks on behalf of users based on instructions, business data, and configured processes. Unlike a traditional AI assistant that waits for a user to ask a question, an agent can perform defined work as part of an ongoing process. In Project Operations, that can mean preparing weekly time entries, converting receipts into expense records, or reviewing submitted transactions against company policies.

Microsoft’s Project Operations agents use components across Dynamics 365, Microsoft Copilot Studio, Power Automate, Dataverse, and related Microsoft services. Administrators configure the environment, permissions, connections, and agent settings before employees begin using them.

Benefits of using AI agents in Project Operations

For project-based organizations, the potential benefits include:

• Less administrative work: Agents can prepare time entries, process receipts, and perform the first review of routine approvals.
• More complete time capture: Automated drafts and missing-time reminders can help employees submit project hours more consistently.
• Faster expense processing: Receipt information can be extracted and organized without employees entering every field manually.
• More focused approvals: Project managers can spend more time reviewing exceptions instead of giving every transaction the same level of attention.
• Timelier project financial data: Getting time and expenses recorded sooner can give project and finance teams more current cost and billing information.
• More time for project delivery: Reducing repetitive administration gives employees and project managers more time to focus on customers, delivery, and project performance.

The goal is not to remove people from project management. It is to reduce repetitive effort while keeping employees and managers involved where review, judgment, or approval matters.

What AI agents are currently available in Dynamics 365 Project Operations?

Microsoft currently has three standard AI agents for Dynamics 365 Project Operations: the Time Entry Agent, Expense Agent, and Approvals Agent. All three are currently classified as production-ready preview capabilities. Organizations can also develop custom agents with Microsoft Copilot Studio when their requirements extend beyond the standard agents.

Time Entry Agent

For many professional services organizations, time entry is necessary for project costing, utilization, customer billing, and financial reporting. However, employees may still wait until the end of the week to reconstruct what they worked on.

The Time Entry Agent is designed to reduce that burden. It can create draft weekly time entries based on sources selected by the user, including:

• Project task assignments
• Booked resource capacity
• Previous week’s time entries

When multiple sources are selected, Microsoft prioritizes project task assignments, followed by booked capacity and previous entries. Users can review and change the drafts before submitting them for approval. The agent can also alert users about missing hours. It can use relevant Outlook calendar information when creating external comments when calendar access is configured.

The business benefit goes beyond saving time on a timesheet. More complete and timely time capture can give project managers and finance teams better information about actual effort, project cost, utilization, and work that may be ready for billing.

Expense Agent

Expense reporting creates another administrative gap between performing project work and recording its true cost.

The Expense Agent uses AI to help process receipts and create expense information. It can extract receipt data, generate expense entries, itemize more complex receipts, and group expenses into reports.

Rather than requiring employees to manually type every value from a receipt, the agent prepares much of the information for them. Employees can review the results before the expense moves through the organization’s normal process. For organizations with consultants or employees submitting a high volume of project expenses, this can reduce administrative work. It can also help get project costs recorded sooner.

Activity Approvals Agent

Project managers often need to approve large numbers of routine transactions. Reviewing every time entry, expense, and material transaction with the same level of attention can create an approval bottleneck.

The Activity Approvals Agent performs an initial review of submitted:

• Time entries
• Expenses
• Material usage

The agent evaluates transactions against policy documents configured by the organization. It then classifies records as Ready for approval or Needs review. Organizations can maintain organization-wide policies. They can also apply different policies to individual projects when necessary. This lets project managers focus first on transactions that appear to require attention instead of treating every submission as an exception.

Organizations can also configure the agent to automatically process records classified as Ready for approval. Approval decisions can affect project costs, billing, reimbursement, and financial controls. Microsoft therefore recommends thoroughly testing the agent before using automatic approval.

Custom AI agents with Microsoft Copilot Studio

The three Microsoft-provided agents address common Project Operations processes, but they will not cover every organization’s requirements.

Microsoft Copilot Studio gives organizations a platform for creating custom agents and connecting them with business data, workflows, and other Microsoft services. Project Operations organizations can use the broader platform to design agents around processes specific to their business.

Potential custom scenarios might include:

• Following up on missing project information
• Preparing a project manager briefing before a status meeting
• Identifying projects that require management attention
• Supporting resource planning or staffing workflows
• Routing project exceptions to the right person
• Automating organization-specific project administration

These are examples of potential custom use cases, not standard Project Operations capabilities.

Before building a custom agent, start with the business process. Define the trigger, information the agent needs, actions it can take, exceptions it should escalate, and how success will be measured. Customization is most useful when it solves a clear gap rather than duplicating functionality Microsoft already provides.

How do AI agents fit into the Project Operations lifecycle?

The value of these agents becomes clearer when they are viewed as part of the broader project lifecycle:

Project planning → Work performed → Time and expenses captured → Transactions reviewed → Approval → Billing and financial reporting

The Time Entry Agent helps prepare time records, while the Expense Agent converts receipts into expense information. The Approvals Agent reviews submitted transactions and identifies exceptions. The resulting project data can then continue through Project Operations processes for costing, financial tracking, and billing.

What is the difference between Copilot and AI agents in Project Operations?

Copilot and AI agents play distinct roles within Project Operations, even though they often work together. Copilot focuses on assisting users through natural language interaction, helping them understand data, generate content, and navigate the system more efficiently. AI agents, on the other hand, are designed to handle work in the background by monitoring data and executing tasks with limited human involvement. This separation allows Project Operations to support both interactive guidance and operational automation.

How much do AI agents in Dynamics 365 Project Operations cost?

Microsoft’s prebuilt Project Operations agents require Copilot Credits, which are sold separately. The Time Entry Agent, Expense Agent, and Approvals Agent do not each have a published flat monthly price. For organizations in the United States, Microsoft currently offers several ways to purchase the credits used by agents:

An active Azure subscription is required to use the pay-as-you-go option for Dynamics 365 agents. The amount an organization ultimately spends depends on how much work its agents perform. Microsoft bases Copilot Credit consumption on the actions and AI processing used. As a result, a simple action does not necessarily consume the same amount as a more complex workflow.

Custom Copilot Studio agents also use Copilot Credits. More complex agents may consume more because they can use generative AI, retrieve information, invoke tools, perform actions, or run multi-step processes. For a detailed breakdown of consumption-based AI pricing, read our guide to Microsoft Copilot Credits.

What should you plan before enabling AI agents in Project Operations?

Microsoft’s standard agents can reduce repetitive work, but an agent will only be as useful as the process and data behind it. Before deployment, focus on a few practical areas.

• Start with a measurable problem. Identify what you want to improve, such as late time entry, slow expense reporting, or an approval backlog.
• Review your Project Operations data. The Time Entry Agent depends on information such as project task assignments, resource bookings, and previous entries.
• Document approval policies. If you plan to use the Approvals Agent, make sure your time, expense, and material policies are clear enough for consistent application.
• Define human review points. Decide where employees, project managers, and finance teams need to verify AI-generated information or handle exceptions.
• Review security and connections. Agents can rely on Dataverse, Power Automate, Copilot Studio, Outlook, Teams, and other services depending on the scenario.
• Pilot before expanding. Start with a controlled group of users or projects. Measure accuracy, administrative time, exceptions, adoption, and AI consumption.

Organizations planning a broader deployment can also use the Dynamics 365 Project Operations implementation guide. It covers the data, process, security, integration, and training decisions that affect both the core system and AI capabilities.

How to get more value from AI agents in Project Operations

Microsoft’s standard agents give organizations a practical starting point for agent-based automation. Begin with a focused process and measure how the agent performs. Expand only when the results support a broader rollout.

Rand Group can help organizations:

• Identify processes where AI could reduce repetitive work
• Design custom agents beyond Microsoft’s standard capabilities
• Connect agents with Project Operations and other business systems
• Improve the project data and workflows agents depend on
• Establish security, governance, and cost controls
• Measure usage and refine agents after deployment

An organization might, for example, start with the Time Entry Agent. It could later identify a need for a custom agent that brings together project status, resources, financial data, and open actions for project managers.

Rand Group works across Dynamics 365, Power Platform, Copilot Studio, and business process design. For organizations planning a broader AI initiative, our Microsoft Copilot consulting services cover strategy, governance, custom agents, deployment, and adoption.

Frequently asked questions

Put AI agents to work in Dynamics 365 Project Operations

AI agents give Project Operations users a practical way to reduce repetitive work across time entry, expenses, and approvals. They also keep people involved in decisions that require judgment. Organizations can start with Microsoft’s prebuilt agents, measure the impact, and use Copilot Studio when a clear need exists for more specialized automation.

The best starting point is a specific process where administrative work slows employees, project managers, or finance teams down. If you want help identifying that opportunity, building a custom agent, or connecting AI with your broader Project Operations strategy, contact Rand Group to discuss your goals.