Morphe: the AI agent that captures and structures your employees’ critical know-how

More than 10,000 workers retire every day in North America, taking their expertise with them. To address the challenge of retaining this knowledge within organizations, Vooban’s team has launched Morphe, an AI agent that interacts with your employees to build a structured knowledge base. Culturemania attended the unveiling of this new product, presented on May 7 at the Impact IA event, held at the Grand Quay of the Port of Montreal.

How does it work?

First, the employee sits down to speak with Morphe during a half-hour session, in which the conversational agent asks targeted questions about their role, decisions, and processes, while following predefined rules and learned frameworks. During the interactive interview, Morphe builds a connected knowledge base that your team can engage with through the multi-agent system deployed within your organization.

“The manager defines the topics to be covered in advance, in collaboration with the employee.”

But how do you determine which critical elements to focus on? “The manager defines the topics to be covered in advance, in collaboration with the employee. If the employee explains two things without a clear connection between them, the AI agent will detect this gap and ask follow-up questions to fill in the missing information, because it is often harder for employees to know what they’ve left out,” explained Carl Chouinard, Product Lead at Vooban.

An observation echoed by Geoffrey Bernard, VP of Marketing, Data Science, and Consumer Insights at La Presse—presented at the event as a pioneer in adopting the multi-agent shift: “We realized that for some, it was a major challenge to surface the substance that gives value to their expertise—to extract that critical value from people’s minds and turn it into autonomous agents.”

“We realized that for some, it was a major challenge (…) to extract that critical value from people’s minds and turn it into autonomous agents.”

Morphe is designed to integrate with a company’s existing technological environment. In addition to text-based input, developments are underway to enable visual and audio capture directly in industrial settings via mobile devices. The platform relies on a human feedback loop to validate the accuracy of the data collected and to ensure information sovereignty.

In response to concerns legitimately raised by participants, Vooban maintains that while many tasks can be automated depending on each role, human judgment remains essential. The agent may make certain decisions, but it is not always capable of handling new situations that require human analysis. While the tool may pave the way for automating some roles, the emphasis is placed on automating specific tasks rather than systematically replacing individuals as a whole. To prevent the agent from generating incorrect information or autonomously “filling in the gaps,” the system is built around a continuous feedback loop with the user. It is therefore the human who validates the agent’s understanding and ensures the accuracy of the captured knowledge before it is made available to the team. Ultimately, the primary benefit is to counter the rapid obsolescence of technical documentation by enabling continuous and easily updatable knowledge through Morphe’s communication interface.