Prism is a conversational agent that helps employees better understand their needs and target the exact Opportunity for those needs.
The introduction of an AI agent into your platform may raise doubts and questions. Here are our answers.
How the page works
The page is accessible via right profiles. Activating the feature also adds a new permanent section to the guide.
A temporary card is also displayed on the homepage. This disappears once the employee has accessed the agent’s dedicated tab (whether via this card or another entry point).
On the agent’s dedicated page, employees have access to:
a video to help them understand how the agent works
a section about data security
How does a conversation work?
The employee can directly ask a question in the chat or use one of the three starter prompts available. The agent conducts the conversation in the employee’s current platform language but can suggest Opportunities in all languages. The agent compiles the information it has about the employee, asks for more details about their background and/or Wish for development if needed, until it can select sufficiently relevant Opportunities.
Conversations are independent from one another. The agent does not retain requests from previous conversations. The agent also does not update the profile based on new information provided by the employee.
The employee can access their conversation history. If Opportunities previously suggested are no longer valid (because they have been archived or no longer meet Matching rules), this will be explicitly indicated.
Information used by the agent
From the profile
The Matching agent uses the following data from the employee’s profile:
Their Job, domain, family
Their skills
Their professional experience, assignments, professional training, and education
Their location preferences
Their bio (free text on the profile)
From Opportunities
The Matching agent uses the following data from Opportunities:
Title
Description
Associated skills
Opportunities suggested by the agent
For now, the agent only suggests open positions. It may suggest developing certain skills through training, but it will not directly push training Opportunities. If the agent does not find Opportunities that exactly match the employee’s criteria but finds close matches, it may indicate this—the employee can then choose whether to receive these suggestions as well. Each suggested Opportunity includes an explanation so the employee can understand the context and why it matches their needs.
The Opportunities suggested by the agent are the same as those suggested through standard recommendations. In particular:
Matching rules are always applied
Previously expressed rejection reasons from the employee are always considered
Opportunities must not have already received feedback (positive or negative) from the employee to be suggested
Opportunities whose type has been configured not to generate matches or not to appear on Opportunity pages cannot be suggested
Employees whose administrative status prevents them from having matches will not see the Prism tab in their interface.