Setting up AI assignment prediction

DMF 090188 ADMINISTRATOR SETUP ASSIGNMENT PREDICTION

Important

This feature requires the feature flag Portail Collaborateur domain > Facilitation Saisie Temps label.

The CONNECT_CALENDRIER management rule must be enabled and the associated setup completed (see CONNECT_CALENDRIER).

Assignment prediction can be set up to enable a large language model (LLM) to automatically determine and specify the assignments for timesheet entry suggestions generated from the Web Portal > Employee > Timesheets & schedules. Assignments are predicted based on previously completed and approved suggestions.

The assignment prediction setup can be accessed via the Employee tab on the left of the Settings Portal, then on Calendar > Assignment prediction.

Setups for the various companies are available from the tabs, at the top of the assignment prediction setup screen. By default, the ***** tab is the only one displayed.

The setup done on the ***** company is applied by default to all other available companies in Akuiteo, unless there is a specific setup for a company. A user who logs into the Web Portal has access to the specific setup for their company first or, if there is no specific setup, to the setup of the ***** company.

1 From the assignment prediction setup screen, click on the + tab at the top.
2 In the window, select a Company from the drop-down list. The help list displays all the companies set up in Akuiteo that do not yet have a specific setup.
3 Click on Validate.
4 Specify the LLM provider from the drop-down list. Only the OpenAI LLM value is available.
5 Click on Save at the bottom of the screen.

A new tab is added for the specific setup of the selected company.

1 From the assignment prediction setup screen, click on the desired company's tab at the top.
2 Click on Delete setup at the top right of the screen, then confirm the deletion.

The specific setup is deleted. The setup used for this company will now be the ***** company setup.

Important

To implement assignment prediction in your organization, contact your project manager at Akuiteo or subscribe independently:.

  • either through OpenAI;

  • or by deploying an OpenAI LLM on Microsoft Azure.

The LLM must be GPT-5 mini.

To use assignment prediction, you must first set up the connection to a large language model (LLM).

1 From the assignment prediction setup screen, click on the desired company's tab at the top.
2 Enter the information:
Field Description
Enable Check the box to enable assignment prediction on the company.
LLM provider

Select the LLM provider from the drop-down list.

Only the OpenAI LLM value is available.
LLM URL if applicable

Specify the LLM endpoint URL if the model is deployed on Microsoft Azure.

Leave this field blank if you are using an OpenAI-hosted LLM.
3 Click on Save at the bottom of the screen.

The LLM is set up on the company.

The LLM API key must also be specified (see Adding the API key associated with the LLM).

Important

On the ***** company, the API key can only be added or modified by a super-administrator.

Prerequisite: On-going modifications on the setup screen must be saved.

Once the LLM is set up, the corresponding API key provided by the LLM host or Akuiteo must be specified in Akuiteo to establish the connection to the LLM. The API key is the unique identifier that authenticates and secures requests to the LLM.

1 From the assignment prediction setup screen, click on the desired company's tab at the top.
2 Click on Add API key to the right of the API key field.
3 Specify the API key provided by the LLM host.
4 Click on Test API key to ensure the key is valid.
5 Click on Save API key.

The API key is added.

From the assignment prediction setup screen, click on the desired company's tab then modify the desired fields and click on Save.

Important

On the ***** company, the API key can only be added or modified by a super-administrator.

Prerequisite: On-going modifications on the setup screen must be saved.

1 From the assignment prediction setup screen, click on the desired company's tab at the top.
2 Click on Modify API key to the right of the API key field.
3 Specify the API key provided by the LLM host.
4 Click on Test API key to ensure the key is valid.
5 Click on Save API key.

The API key is modified.