Lydon Solutions

Construction Project Management Software Solutions

AI Solutions

How the New Copilot Project Manager AI Agent Benefits Construction Scheduling

AI Solutions |

The new Project Manager AI agent is now available in preview for Planner Premium if you have a Copilot license. So, how does a Project Manager AI benefit a construction scheduler? Where are the gaps, and does it make sense to pay for the additional Copilot subscription fee to get access to the agent? Check out this post to find out more.

At the time of writing this post, the new Copilot Project Manager AI agent for Planner is still in preview and only available from Teams. You cannot access it directly from the Planner app, which is a little weird. To use the new agent, you will need a Planner Premium and Microsoft 365 Copilot paid subscription.

What can I do with the Project Manager AI Agent?

Updates to Create a Plan

There is a new plan called Plan with Project Manager, which adds a Project Manager tab in Planner for the Copilot agent. Project Manager adds the following features:

Create a Goal

With a simple prompt, the Project Manager can build out a project plan from scratch. You can enter a description of the plan in the prompt and upload associated files. The Project Manager will do the rest.

For the purposes of this post, I made a basic plan using the following prompt: “Excavate 200 cy of soil contaminated with hydrocarbons and replace with clean backfill.” That is probably not the most descriptive or realistic, but it was a starting point.

The Project Manager thought about it for a few seconds and then built out a list of tasks, elaborating on the details of each task within a new Loop component. The AI added finish-to-start logic between tasks, assigned a default one-day duration for each task, and provided a quick view of upcoming tasks. Once task owners are assigned, the Project Manager automatically sends notifications via email and Microsoft Teams.

For a new scheduler or one that provides remote support, the ability of the Project Manager AI agent to build a schedule based on a prompt and previous files is a powerful feature. But it will require a second set of eyes to ensure it’s up to snuff.

Assign tasks to the Project Manager

Before leveraging the Project Manager, you will want to scrub the timeline view (Gantt chart) to ensure tasks have the correct start and finish dates, durations, and logic. The Project Manager is not perfect, and you will need to review its output. Don’t assume it does everything right.

The magic of the agent lies in assigning tasks directly to the Project Manager. This is where it gets interesting.

When you assign tasks to the Project Manager, it will analyze the task information and reach out with additional questions. I added a new task called “Implement a PMIS.” It prompted me to help and asked me to answer several related questions.

Once I answered and clicked the regenerate button to run the Project Manager agent, it built out details of the task, including:

  • Steps to implement the PMIS.
  • Steps to build a Power BI report to analyze documents.
  • Approaches to train users on the system.
  • Sample code for a document upload notification in Power Automate in ~ 30 different computer languages.
  • Samples of metrics to monitor and manage performance.

I wasn’t expecting the extent of the analysis the Project Manager provided. It took my feedback to its questions and further defined the tasks to what exactly needed to happen within the Loop component.

For a construction project, schedule development involves a detailed analysis of each task to understand what is in scope and what it means for a task to be completed. The Project Manager AI agent proved helpful for further defining the specifics of the task and putting a box around the defined scope of the task.

Copilot in Whiteboard

Whiteboard is a great tool for brainstorming ideas. Many construction organizations aren’t aware that it is included in their Microsoft 365 subscription. The Project Manager AI agent can convert Whiteboard notes to tasks and add them to your plan. That’s a great feature for taking concepts and connecting them to actionable items.

For a scheduler, this feature has a lot of potential. Back in the day, construction project planning sessions involved team members identifying tasks, writing them on physical Post-It notes, and arranging them in sequence on a physical whiteboard to define all of the tasks that needed to be completed, who was responsible, and when. This new whiteboard feature has the potential to make this effort more efficient and virtual.

There is a lot more to Project Manager, and Microsoft just released a new Status Report feature. Keep an eye out for part 2 of this post.

As always, if you need help with Microsoft 365 for your construction projects, you can request a free one-hour consultation below, or if you need an enterprise construction management platform for Microsoft 365, check out Construction Viz here.

Get a Free Consultation

Copilot and Planner Premium – R.I.P. Project Schedulers?

AI Solutions |

Still on the fence about whether to jump in and pay for Planner Premium (formerly Project for the Web)? Microsoft has recently integrated Copilot AI into Planner Premium. Find out how Copilot can take scheduling in Planner Premium to the next level.

There are two Microsoft project web applications: Project Online, which has been around since 2016, and Planner Premium (Project for the Web). While Planner Premium might lack some features of Project Online, it is the future of Microsoft Project. With the addition of Copilot and the upcoming Project Manager Agent, Planner Premium is turning into a powerful tool.

So, you might wonder what AI could do for scheduling. I also asked that way back in 2016 when bots were picking up steam. While there are aspects of a scheduler role that cannot be done simply by AI, I can say that we are knocking on the door of having a scheduling AI agent capable of building, analyzing, and reporting on project schedules.

Read on to see how our testing of Copilot to assist with construction project management turned out.

Using Copilot for Construction Project Management

Copilot works by using prompts. The better the prompts, the better the results (assuming the AI is up to snuff). For the purposes of our analysis, we used the out-of-the-box Copilot, which had no agents.

The following are some of the prompts that returned some fascinating results.

Prompt: Build a plan for an excavation construction project.

Result: Copilot returned a list of grouped tasks based on a typical excavation project. Was it perfect? No. While it created the tasks and the associated goals, it did not add dates, durations, or logic.

Prompt: Create a task.

Result:  Copilot creates a task, but again, it has no dates or durations.

Prompt: Delete all or a task.

Result: Copilot will delete one or more tasks, but the goals remain.

Prompt: What is the critical path?

Result: While you can filter a schedule by critical path, the prompt will display the critical path tasks.

Prompt: What if a task is delayed by X days?

Result: This is pretty slick. It did a schedule scenario and provided the impact of the delay on each activity and the end date of the project.

Prompt: Give the status report for the project.

Result: Copilot will provide a formal response that includes the project manager, start date, finish date, duration, overall percent complete, effort, status of each task, and goals. This result could be huge as schedule analysis can take days to weeks in some cases, and Copilot did this in seconds. From there, you could copy and paste it into an email or a Copilot page and send it off to the project team for your weekly schedule review.

Prompt: How can we expedite the procurement process?

Result: This was a prompt provided by Copilot that we clicked. Copilot returned a mitigation approach that includes a general list of topics with explanations to consider in order to get the procurement process back on schedule.

Prompt: What activities are delayed against the baseline?

Result: We first created a baseline and then progressed the schedule. When we entered this prompt, Copilot summarized which activities have been delayed as well as the impact on the overall schedule compared to the baseline.

Evaluating Copilot for Construction Project Management

Our testing reveals some shortcomings and caveats in using Copilot to help provide scheduling support on construction projects. However, it also shows how fast this technology is improving.

Obvious Gaps

  • Copilot does not include start date, end date, duration, or schedule logic when adding tasks.
  • You cannot prompt Copilot to update the percent complete, schedule logic, or status of an existing task.
  • You occasionally have to close out of Copilot and reopen to get a prompt to run.

Additional Considerations

Planner Premium saves data to the Dataverse just like Dynamics and Power Apps. This integration opens many options for building custom Copilot agents to perform specific tasks for your project.

Lydon Solutions has been building custom solutions using Power Apps and Power Pages to augment Planner Premium, such as a Schedule of Value application for contractors to upload an Excel SOV for approval.

If you would like to know more about Planner Premium or our catalog of solutions, you can request a free one-hour consultation here.

Get a Free Consultation

AI Invoice Processing Solution for Microsoft 365

AI Solutions |

Lydon Solutions is actively developing AI models to help construction project teams be more efficient. Our goal is to use AI to solve real-life construction problems using our CloverAI suite of AI models for Microsoft 365.

One of our latest models is an AI document processor. It can process any file, such as RFIs, invoices, or submittals. This AI model extracts data from PDF email attachments and inserts the metadata into SharePoint document library columns for reporting and approvals.

 Read on to find out how we make your life easier by using AI.

AI for Construction Project Management

Artificial Intelligence is all the rage. Most people are familiar with Open AI and now Microsoft Copilot but are still unsure how it will solve their specific challenges. The promise of AI is incredible and will forever change our industry, but how will it solve the needs of your construction projects today?

Lydon Solutions has been investing in developing AI models to make managing your construction projects easier using the tools already available in your Microsoft 365 tenant.

This article will discuss how we use AI to solve an invoice processing challenge and how we can use this model to help your organization.

CV-Co-Pilot

ChatGPT (Microsoft CoPilot) for Construction Viz

AI Copilot (ChatGPT) now has features across the entire Microsoft 365 and Windows ecosystem.
Read more on Construction Viz >

AI Invoice Processor Solution Use Case

Problem statement

How can we make invoice receipt, entry, and approval as efficient as possible using AI in Microsoft 365?

Approach

We brought together several available technologies in a client’s 365 tenant to solve this challenge.

AI Invoice Processing Solution Detail

Below are the technologies and steps used to deliver the solution.

Outlook email:

  • We started by considering that most invoices are received via email from contractors because they likely won’t have guest access to your Microsoft 365 tenant.
  • This could pose a challenge, but by leveraging a Microsoft 365 group email, we can extract attachments using a workflow when vendors send invoices to this email address.

Power Automate workflows:

  • Receipt Workflow.
    • We built a receipt workflow that extracts PDF invoice attachments from emails and saves them to a SharePoint document library with the date received, sender, and email subject line.
    • This step addresses the issue of lost invoices by consolidating all invoices into one location.
  • AI Workflow.
    • We created and trained an AI model to read the PDF invoices and extract metadata, including company and amount, amongst other fields.
    • This metadata is then used with the document stored from the Receipt Workflow, eliminating the need to enter metadata or move invoices into folders manually.
  • Once the AI workflow extracts the data successfully, it sends confirmation emails to the sender and the person tasked with approval.  
  • Additionally, the Microsoft Teams approvals app can be updated if needed. 

SharePoint site and document libraries:

  • We use a SharePoint site as an invoice repository and portal for invoice approval and management.
  • Once an invoice is approved, a final workflow triggers and sends a confirmation email to the teams.

Check out this video to see the whole process in action:

Learn About Using AI for Construction Project Management

If you are interested in this AI model or need other models for your construction projects, complete the form below to request a free one-hour consultation.

If you are looking for a Microsoft 365 enterprise construction management solution that also leverages these AI models, you can request a free demo of Construction Viz here.

Get a Free Consultation