Supper Skills are reusable workflows for data analysis. They give Supper's AI agent the instructions, business context, calculations, and success criteria it needs to complete recurring analytical work—such as customer health scoring, churn analysis, or pipeline risk reviews—without requiring a team to rebuild the process from scratch each time.
Skills are built for work that is too involved for a single question, but too common to keep recreating in spreadsheets, prompt chains, or one-off analyses. A team can define a Skill once, then use it whenever the same type of analysis comes up again.
What a Supper Skill includes
Each Skill describes a specific job for the agent. Depending on the workflow, it can include:
- Required inputs from the user
- The data and business concepts the agent should use
- Instructions for retrieval, analysis, and calculation
- Algorithms or scoring logic
- Tools the agent needs to use
- Rules for investigating missing or unexpected information
- The format and criteria for a finished result
The result is not a fixed macro. A Skill can tell the agent what it needs to accomplish while leaving room to investigate the data and make decisions along the way.
For example, a customer health-score Skill might instruct Supper to pull product usage, account activity, support data, and commercial information; calculate the relevant factors; and return a score with the signals that explain it. The exact customer data will change from one run to the next. The method does not have to.
Where Skills are useful
Skills are meant for recurring analysis that has multiple steps or requires consistent judgment. Common examples include:
| Workflow | What the Skill can do |
|---|---|
| Customer health scoring | Combine usage, commercial activity, support signals, and engagement into a health score. |
| Churn analysis | Identify churned or at-risk customers and surface the signals behind the result. |
| Pipeline risk review | Evaluate pipeline records against the criteria a revenue team uses to spot risk. |
| Variance investigation | Guide a finance team through the data and calculations needed to explain a material change. |
These workflows often begin as exploratory work. An analyst asks a question, follows the evidence, refines the logic, and arrives at an answer the team trusts. A Skill makes that work available after the analysis is done.
Skills use the Supper semantic model
Skills sit on top of Supper's semantic model. That matters because a reusable workflow is only useful if the agent understands the data and the business language behind it.
The semantic model provides that foundation: tables and columns show where data lives; descriptions explain what it represents; and Terms define reusable business concepts, calculations, and logic. Skills use those building blocks to capture a process.
| Semantic-model layer | What it provides |
|---|---|
| Tables, columns, and relationships | The underlying data structure |
| Descriptions | Context about the data |
| Terms | Shared definitions, calculations, and business logic |
| Skills | A reusable process that applies those definitions and calculations |
For instance, a Term may define what counts as an active customer. A Skill can use that Term as one input in a broader customer health workflow. Keeping both in the semantic model means the agent can use the same definition wherever it is needed.
A Skill is not a saved prompt
A saved prompt repeats an instruction. A Skill captures the work behind an outcome.
That difference is most visible when the workflow needs to adapt. A prompt might say, "Find at-risk customers." A Skill can define which signals indicate risk, what data to check, which calculations to apply, what exceptions to investigate, and what the final analysis should include.
Skills can also reference other Skills, semantic-model context, algorithms, and specialized tools. They can include reasoning loops for work that requires follow-up investigation rather than a single pass through a fixed sequence.
Skills work with Supper's existing tools and controls
When a user runs a Skill, the agent uses Supper's existing capabilities to do the work. That may include retrieving data with SQL, manipulating results with Python, running calculations, or creating a chart when a chart is the right output.
The Skill provides the process. Supper's agent uses the appropriate tools to execute it.
The same permissions and entitlements continue to apply. A Skill does not give anyone access to data they could not otherwise access; it guides the agent through analysis using the access the user already has.
Skills are also versioned as part of the semantic model. Teams can review how a workflow has changed and return to previous versions when necessary. That is useful when a health-score definition changes, a risk model gains a new input, or a team updates its process after learning something new.
Creating and improving Skills
Teams can create Skills manually or work with a dedicated Skills agent. In practice, a Skill can start with work a team has already done: an analyst works through a problem, the team agrees on an approach, and that approach becomes a reusable workflow.
Some Skills will be tightly specified, with required inputs and clear steps. Others can be more open-ended, giving the agent the business objective and the guidance it needs to carry out a sound investigation. If a required input is missing, the agent can ask for it before proceeding.
Skills give teams a practical way to preserve analytical knowledge that would otherwise live in a spreadsheet, a dashboard, a chain of prompts, or one person's memory. When a process is worth repeating, it should be possible to save it where the data work happens.
Frequently asked questions
What are Supper Skills?
Supper Skills are reusable workflows for multi-step data analysis. They give Supper's AI agent the business context, instructions, calculations, tools, and success criteria needed to complete recurring analytical tasks.
Are Supper Skills just saved prompts?
No. Saved prompts reuse instructions. Skills capture an analytical process, including the inputs, logic, tools, checks, and expected output required to produce a reliable result.
What can I use Supper Skills for?
Supper Skills can support customer health scoring, churn analysis, pipeline risk reviews, variance investigations, and other recurring work that requires multiple data pulls, calculations, or investigation.
How do Skills work with Supper's semantic model?
Skills use the context in the semantic model—such as tables, descriptions, Terms, and business logic—to carry out a reusable process. The semantic model explains the business and its data; a Skill explains how to complete a specific analysis using that foundation.
Do Skills change data access permissions?
No. Skills use Supper's existing tools and respect the user's existing permissions and entitlements. They do not grant access to data a user could not otherwise access.