Skills overview

A skill is a specific, observable capability someone can develop, for example “Gives actionable feedback” or “Closes a support ticket cleanly.” Each skill defines how it’s assessed and what counts as mastery, so progress is measured consistently across your team.

Skill: the unit of management

Admire treats the skill as the unit of management. Almost everything a manager cares about ladders up to skills: a job description is really the set of skills a role requires, a performance review is a snapshot of skills demonstrated, and a development plan is a path to the next skill. Anchoring on skills works because skills:

• Drive results: they directly affect whether goals are met.
• Are in people’s control: unlike outcomes, someone can actually work on a skill.
• Align everyone: individuals, managers, and AI share one clear definition of what good looks like.
• Standardize and automate: because skills are defined consistently, AI can track them, coach on them, and handle the legwork.

So the things a manager already works with are really views of the same underlying skills:

flowchart LR
  Skill(("Skill"))

  Skill -- "increases success on" --> Goals["Goals"]
  Skill -- "makes up" --> JobDescription["Job description"]
  Skill -- "is demonstrated in" --> Review["Performance review"]
  Skill -- "is the next step in a" --> Plan["Development plan"]

What makes up a skill

• A description of what the skill is and what good looks like. This captures the what, not the how. How to actually carry out the skill belongs in a linked playbook page, so an assessment can point straight to the guidance for closing a gap.
• An assessment type: how you score it (see Assessment types).
• A mastery threshold: the score (0–100) to reach, and the number of consecutive sessions at or above it required to count as mastered. This is always at least 1; requiring more than one consecutive session makes mastery reflect consistency, not a one-off.

Skill trees

Skills can be organized hierarchically, with broader skills made up of narrower ones. A composite skill can roll up its children, letting you track both the big-picture capability and the specific behaviors underneath it.

Recommended structure: 1 – 5 – 25

A simple three-level shape works well for most teams:

• 1, Skillset (root): the broad capability, e.g. “Client communication.”
• 5, Skill (child): the specific skills within the skillset, typically graded on a gradient (e.g. a 1–5 rating). This is the level you usually assess.
• 25, Behaviour (grandchild): concrete, binary (yes/no) behaviors that make up each skill, a checklist that makes assessing the parent skill easier and more consistent.

The behaviour level is optional. It turns a judgment call into a checklist, but an assessor can always assess the skill directly if they prefer.

Here is the shape applied to a real tree from the Skills Vault, the “Apply AI Tools Effectively in Daily Work” skillset. Notice the behaviour count varies per skill: some have several, some have none.

Category	Name	Assessment type
Skillset	Apply AI Tools Effectively in Daily Work	Composite
Skill	Structure Effective AI Requests	Rating
Behaviour	↳ Provide relevant context and constraints	Binary
Behaviour	↳ State the task and desired output clearly	Binary
Behaviour	↳ Include an example of the result you want	Binary
Behaviour	↳ Specify the format and length	Binary
Skill	Iterate, Refine, and Co-Create with AI	Rating
Behaviour	↳ Critique and redirect weak responses	Binary
Behaviour	↳ Combine AI output with your own expertise	Binary
Skill	Redesign Personal Workflows Around AI	Rating
Skill	Communicate AI-Assisted Work to Stakeholders	Rating
Skill	Adapt Continuously as AI Tools Evolve	Rating
Behaviour	↳ Evaluate new AI tools against workflow needs	Binary

Each level plays a different role:

• Skillset is for organization. It groups related skills under one capability and isn’t assessed directly; its score rolls up from the skills beneath it.
• Skill is the level you usually assess directly. Because it’s rating-based, a coach uses their judgement to decide where the person lands.
• Behaviour aids skill assessment by grading in a more yes/no manner. When a coach is unsure how to rate a skill, they can assess its behaviours to guide them and build up toward the skill assessment. Behaviours are optional, so a skill can have none.

Creating skills

You can build skills three ways. We recommend them in this order:

1. Use AI over MCP (recommended): connect an AI tool and let it guide skill creation. It understands Admire’s structure and asks the right questions, so it’s the fastest way to a well-formed skill.
2. Import from the Skills Vault and adjust: browse the Vault by category, import a ready-made skillset, and tailor it to your context.
3. Create directly: once you’re comfortable with how skills are structured, define one by hand.

You can edit a skill’s description and settings at any time, and disable a skill that’s no longer relevant without losing its history.

Browsing the Skills Vault

The Skills Vault is a storefront of expert-built skillsets, what good looks like, ready to import and tailor to your team. The home page leads with a featured skillset and then rows of skillsets grouped by category, so you can scan by the outcome you care about instead of hunting through a flat list.

The Skills Vault home, with a featured skillset up top and skillsets grouped by category.

Importing brings the whole skillset into your organization (its skills, their behaviours, and any linked playbook pages and AI coaches), where you can adjust it.

Browse skillsets and filter by category, importing any skillset with one click.