Core concepts
The building blocks of Sleak and how they fit together.
Sleak is built around three core concepts:
- A conversation type sets the standard for one kind of conversation. Its scorecard says what a good conversation looks like.
- An AI-Persona is the counterpart employees practise that conversation with.
- A program brings personas and missions together in a training path with a due date.
Everything else in Sleak builds on these three.
The three core concepts
Conversation types and scorecards
A conversation type describes one kind of conversation employees train, for example a discovery call, a price negotiation or a feedback talk. It sets the standard for that conversation:
- The goal says what counts as success.
- The scorecard holds the phases and the criteria. The phases are the steps of the conversation. The criteria describe what a good conversation looks like in each phase.
Sleak scores every conversation against the scorecard of its conversation type. In practice, the conversation type is your company's standard, written down.
Read more in Conversation Types.
AI-Personas
An AI-Persona is the counterpart employees practise with. It is one person in one concrete situation, for example a purchasing manager who pushes for a discount.
A persona has:
- exactly one conversation type, which decides how the conversation is scored,
- a briefing that the learner reads before the conversation,
- a difficulty: easy, medium or hard,
- a voice, or a text chat instead of voice.
A new persona is private. It becomes available once you publish it and share it with the workspace, with teams or with people.
Read more in AI-Personas and AI-Personas Management.
Programs
A program is a structured training path that you assign to people or teams. It has a due date and is made of levels, which unlock one after the other.
Each level contains personas and related missions, for example "three conversations with a score above 70". A learner completes a mission by practising with the personas in that level.
Use a program when employees should reach a defined standard by a certain date.
Read more in Programs and Program Management.
Conversation type, persona and program
An example with two conversation types, three personas and one program. The program picks personas from both conversation types.
Conversation typeNeeds analysis call
Personas
- IT manager at a mid-sized companyEasy
- Managing director with little timeMedium
Conversation typePrice negotiation
Personas
- Purchasing manager who pushes for a discountHard
ProgramSales onboarding
Level 1
MissionsIT manager at a mid-sized company
Needs analysis call
Three conversations with a score above 70
Level 2
MissionsManaging director with little time
Needs analysis call
Two conversations with a score above 75
Level 3
AssessmentPurchasing manager who pushes for a discount
Price negotiation
One conversation with a score above 80, at most 3 attempts
- A conversation type can have many personas.
- A persona belongs to exactly one conversation type. It decides how the persona’s conversations are scored.
- A program uses personas from any conversation type.
Training in Sleak
Conversations
A conversation is one practice session between a learner and a persona. Afterwards, Sleak scores the conversation from 0 to 100 and gives feedback on every scorecard criterion. The transcript records what was said.
You can also import real conversations. Sleak scores them the same way.
Read more in Conversations.
How a conversation is scored
An example with three criteria from the conversation type Needs analysis call.
AI-Persona
Plays the counterpart in the situation.
Conversation type
Provides the goal, phases and criteria.
Conversation
The learner talks or chats with the persona.
Sleak checks every criterion
- Asks open questions
- Asks about the budget
- Agrees on the next step
Result: A score from 0 to 100 and feedback on every criterion.
Feedback
The learner sees for each criterion what went well and what to improve.
Program
The conversation counts for every mission it fulfils.
Analytics
Managers see how scores develop across teams.
Practice and assessment
Sleak knows two kinds of conversations with a persona:
- Practice is open. Learners practise as often as they like and learn from the feedback.
- An assessment is part of a program. It has a limited number of attempts and no coach. It shows whether a person has proven the standard.
Knowledge sessions
A knowledge session is the reverse of a practice conversation. The coach gives a presentation by voice, asks questions and checks which learning goals the learner has reached. Use it to teach something new, for example a new product or a price change.
Sleak does not score or save knowledge sessions.
Read more in Knowledge and Presentations.
AI Coach
The AI Coach is a chat that helps employees prepare and improve their conversations. It knows their results in Sleak, searches your company's documents and can search the web.
Read more in AI Coach.
Analytics
Analytics shows how much your teams train, how their scores develop and where their strengths and weaknesses are. You can filter by team, person, persona and more.
Your compliance settings decide whether managers see real names or aliases.
Read more in Analytics and Compliance.
Workspace and people
Workspace
A workspace holds everything that belongs to one company: users, teams, personas, conversation types, programs and conversations. Each workspace has its own address, for example app.sleak.ai/your-company.
You can belong to several workspaces. Switch between them with Switch workspace in the workspace menu.
Read more in General settings.
Users and roles
Every person in a workspace has a role. The role decides what the person can do:
- Owners and Admins manage users, teams and settings.
- Content Managers build personas, conversation types and programs.
- Members practise and use the training content.
To practise, a person also needs a license.
Teams
Teams group people, for example by department or region. A Teamlead leads a team and sees the conversations of its members.
Teams matter in three places:
- You share personas and conversation types with teams.
- You assign programs to teams. New members get the program automatically.
- You filter analytics by team.
Read more in Teams.
What a workspace holds
People on one side, content on the other. Sharing and assigning connects them.
Workspaceapp.sleak.ai/your-company
People
Owners and Admins
Manage users, teams and settings.
Content Managers
Build conversation types, personas and programs.
Teamleads
Lead a team and see the conversations of its members.
Members
Practise with a license.
Teams
Group people, for example Sales North or Service.
Content
Conversation types
The standard for each kind of conversation.
AI-Personas
The counterpart to practise with.
Programs
Training paths with levels and missions.
New content is private. Others only see it once you share it with the workspace, with teams or with people.
How these concepts fit together
The way to think about Sleak is:
- conversation types set the standard for each kind of conversation,
- personas turn that standard into concrete situations to practise,
- programs add missions, assessments and due dates to the personas,
- conversations are where employees practise and get scored against the standard,
- analytics shows who has proven the standard and who has not yet.
The workspace holds all of this, together with employees and teams.
How the concepts fit together
From top to bottom: from the standard to the results.
Conversation type
Sets what is trained and what makes a conversation good.
AI-Persona
The counterpart in one concrete situation. Belongs to exactly one conversation type.
Program
Builds missions around the personas, in levels with assessments and a due date.
Conversation
A learner talks or chats with a persona and gets a score and feedback.
Analytics
Shows how teams and people train and improve.
Last updated on 3 November 2026