Articles on: Manage Learning Agents
This article is also available in:

Learning Agent: what is it?

Learning Agents are at the core of the Blify experience. This article explains what a Learning Agent is, what it is used for, how Learners experience it day to day, and what you steer as an Admin.


What is a Learning Agent?


A Learning Agent is an AI-powered conversational training agent that trains your employees on a specific topic by talking with them, directly inside their everyday work tools: Microsoft Teams, Slack, WhatsApp, or the web app.


Three things set it apart from a traditional e-learning module:


  • It comes to the Learner instead of waiting for them to log in to a separate platform. Training happens where the work happens.
  • It holds a conversation instead of "just" broadcasting. Learners read, answer, practice, and ask their own questions.
  • It draws on your content. A Learning Agent feeds on the documents you give it and cites its sources, rather than producing generic material.


What is a Learning Agent used for?


You build a Learning Agent for one named, bounded training need, not to cover an entire catalog. The most common use cases among our customers:


  • Onboarding: deliver the fundamentals to new hires and shorten time to autonomy
  • Sales enablement: train sales teams on an offer, a pitch, or this month's real objections
  • Management: build manager skills around feedback, goal setting, and difficult conversations
  • Compliance: embed internal rules, GDPR, cybersecurity
  • AI adoption: help teams understand how to make better use of AI in their business processes, for more efficiency and productivity
  • Product updates: roll out a new feature to a whole population in a few minutes a day


💡 Good to know: several focused agents work better than one agent trying to cover everything. A clear scope means a shorter, more relevant journey and analytics you can actually read.


What the Learner experiences


The first time a Learner meets a Learning Agent, they receive a welcome message explaining how Blify works, then start their journey.


From there, the exchange runs both ways:


  • The agent reaches out to the Learner several times a week, offering the next activity in their journey. It does not wait for a question, because Learners do not always know what they do not know
  • The Learner reaches out to the agent whenever needed: preparing for a one-to-one, revising before a client meeting, looking up a method they have forgotten


Depending on how the agent is configured, activities take several forms: theory, retention questions and quizzes, and role-plays (voice-enabled AI simulation). Content is personalized for each Learner, based in particular on their job title.


💡 Good to know: role-plays run in a dedicated immersive space on the web app. From Microsoft Teams or Slack, Learners get there through a personalized link that signs them in automatically, with no credentials to type.


What a Learning Agent is made of


A Learning Agent rests on 4 building blocks, matching the 4 steps of its configuration in the admin space.


1. Framing. What the agent must achieve, and for whom.


  • Learning objectives: what the Learner should know or be able to do by the end of the program
  • Target audiences: who the agent speaks to, so the level and the wording can be adapted
  • Prerequisites: the knowledge or skills needed to follow the agent in good conditions


2. Resources. Your sources of truth, meaning the reference content the agent relies on: documents (PDF, Word, PowerPoint, SCORM, and so on), links, multimedia files such as videos, images, or audio, and even your own tools. This is what grounds the training in your business reality rather than in generic material.


3. Structure. The skeleton of the journey, across 3 levels, modules, sections, and topics. These are the "rails" your agent runs on. The structure represents every topic that will be covered with your Learners.


4. Configuration. How the agent speaks to the Learner:


  • Tone: the communication style the agent adopts
  • Addressing: the level of formality, formal or informal
  • Capabilities switched on: quizzes, role-plays, before-after assessments


All of these are set up from the Learning Agents page of your admin space. Learn more about configuring a Learning Agent in 4 steps.


What stays under your control


Blify is built to avoid the black-box effect of AI agents: you remain the person in charge of the training, and the AI does the shaping work.


  • You approve the structure. The outline the AI proposes stays editable, module by module, section by section, topic by topic
  • You preview the content. Click a topic and you see an example of what will actually be generated for a Learner
  • You own the sources. The agent relies on the resources you provide and shows where the information comes from
  • You choose the capabilities. Role-plays make sense for interpersonal skills, far less for a regulatory text: it is up to you to switch them on or off


💡 Good to know: a Learning Agent stays editable after it is published. If Learners are already enrolled, Blify warns you before you make your changes.


Only Users with the Admin role can create and configure Learning Agents. Learn more about roles and permissions.

Updated on: 27/07/2026