Workplace Service Delivery – AI Agents September Release
This blog is a deep dive into the new Agentic AI features for the September 2025 release of the Workplace Service Delivery (WSD) Store Apps.
AI for Workplace Service Delivery – Indoor Mapping AI Agent
Persona: Workplace Team
Roles: n_map_core.map_admin
Plugins: AI Agents for Workplace Service Delivery, Workplace Indoor Mapping,
We are excited to announce a new enhancement to our indoor mapping capabilities: AI Assisted source matching for Bulk Imports. This feature is designed to improve the accuracy and efficiency of updating floor files across large scale environments, such as multi-building campuses.
Previously, the bulk import process required an exact match between the uploaded file name and the original source name. Even minor changes, such as appending a date or adding a version label, could prevent the system from recognizing the correct floor, resulting in failed imports and manual intervention.
To address this, an AI Agent has been introduced to analyze existing records and identify close matches, even when file names slightly differ. This allows the flow to proceed with the import process, without requiring manual interaction.
How it works:
To start the bulk import, the Map Admin would need to go to Map Studio, and selects the option to “Bulk Update Maps”
Once the “Import files” option has been clicked, a new floor plan import group will get created.
Within the floor plan import group view, the Map Admin can define the name of the import, enable “Notify User” option to be notified once the import process has completed and enable “Supervised AI execution” to provide the AI a set of steps to follow
After attaching the files with the updates for existing preconfigured floor plans, the Map Admin can click on “Start Import” to begin the process.
There is an option to trigger the AI agent manually if desired, otherwise the AI agent would run automatically
For this example, the file loaded has a different name than the source file of the map that needs to be updated, previously this would stop the import process, and the Map Admin would need to resolve this manually. As of our latest release, an AI agent will handle the error.
After identifying the error, the Agent will look for a source file with a similar name as the one uploaded, once it identifies a close match it will update it for the import process to move forward.
The agent will perform different actions that get logged as work notes as seen in the screen below.
After the error has been resolved, the import process will continue. If no more errors are identified, the import will be completed, and the status will change to “Closed Complete” meaning the updates have been successful.
AI for Workplace Service Delivery – Workplace Cleaning AI Agent
Persona: Workplace Team
Roles: workplace_user, now_assist_panel_role, sn_wsd_case.case_writer
Plugins: Now Assist for Workplace Service Delivery, Workplace Maintenance Management, Workplace Case Management
As part of our September release, we are excited to introduce a second AI Agent to Optimize Cleaning Activity, designed to manage and streamline Workplace Maintenance operations. This enhancement leverages reservation and visitor data to ensure cleaning activities are scheduled efficiently and only when needed to reduce unnecessary work if a space hasn’t been used.
The Optimize Cleaning Activity agentic workflow integrates with the Workplace Maintenance Management application, which allows workplace teams to create maintenance plans based on location and schedule. These plans can generate cases, such as daily or hourly cleaning tasks for common areas. However, until now, these cases were created without consideration for actual space usage.
With the new AI Agent, utilization rate of each space is evaluated based on reservations and visitor traffic to adjust the number of cleaning cases generated. If a space is underutilized, the agent will deactivate excess cases while ensuring at least one cleaning task remains scheduled. Similarly, if utilization exceeds a defined threshold, additional cases will be created to meet increased demand.
Administrators can configure this behavior using three new properties:
- Plan Selection: Specify which maintenance plans should be optimized
- Maximum Utilization Threshold: Define the usage level for when new cases should be added.
- Minimum Utilization Threshold: Set the usage level for when cases should be cancelled..
How it works:
Workplace Managers have access to Workplace Maintenance Management within Workplace Central. Within Preventive Maintenance they can create a maintenance plan. The purpose of maintenance plans is to perform different tasks at specific locations during a set schedule.
The plan requires a start date and end date as well as a set plan type which can either be location or model. When creating a plan by location, all the spaces within it will be added as “Maintenance Items” which refers to what needs to be maintained.
Schedules can be created for every maintenance item, and they can be defined to run on a specific cadence as well as their trigger.
Once a schedule is created, a plan record will get generated. The “Next run time” column provides information on when will the plan be executed next.
Scheduled jobs based on the “Next run time” will create maintenance cases to execute the plan
Once the scheduled job has been executed, the next run time will be updated to the following day (This will depend on the cadence set for the plan, i.e. monthly cadence would mean that the next run time would be a month from the day the scheduled job was executed)
Every time the scheduled job is executed a new maintenance case will be created. The AI Agent will work on optimizing these cases by decreasing the number of maintenance cases if the utilization rate falls below an established. Similarly, the AI agent will add maintenance cases if the utilization rate is above the established limit.
In order to determine the utilization rate, 3 new system properties have been introduced.
The first system property is used to optimize plans. If desired, multiple plans can be optimized at the same time, simply by adding their ID in the description field as shown in the example below.
The second system property will determine the maximum threshold, supporting the AI agent to determine whether more cases need to be added. If the utilization rate goes above 80% new cases will be created to support that plan.
Lastly the third system property introduces is for the lowest utilization threshold. If the utilization rate falls below 40% the AI agent would decrease the number of cases initially created to support a plan.
A scheduled job to optimize cleaning activities has been introduced. When executed it will analyze the cases created for a maintenance plan and try to optimize them.
To verify whether cases have increased or decreased, a new table called “Execution Plan” has been added
The cases that were initially created for the maintenance plans will be displayed in the execution plan table and in order to see whether the case will be optimized or not, it can be selected, and the response will determine the utilization rate. In the example below, we can see that the utilization rate was 100% which means new cases would be created.
The utilization rate will be defined by the number of reservations and visitor registrations for a specific space.
The output of the AI agent will be a set of scheduled generated in alignment with the utilization rate evaluation that was performed.
