Scenarios

A scenario is a practice situation. It pairs a persona with a goal, a set of objections, and the criteria for “the trainee handled this well”. Build scenarios from real calls — the closer to your actual customer base, the more useful the training.

Anatomy of a scenario

Field	What it does
Title	Short and concrete — `Refund objection — gold tier`.
Persona	The customer the trainee will talk to.
Opening line	What the customer says first.
Goal	What the trainee should accomplish.
Objections	Things the AI customer will push back with.
Success criteria	Conditions for “passed”.
Documents	SOPs, scripts, case files the trainee can reference.
Rubric	The scoring criteria — built-in or custom.
Difficulty	`Easy`, `medium`, `hard`. Drives mood drift and objection density.

A worked example

Field	Example
Title	`Refund refusal — out of policy window`
Persona	`Frustrated cardholder — Maya`
Opening	”I want a refund for the espresso machine. It’s been three weeks and the warranty is gone.”
Goal	Empathize, explain the policy, offer a partial credit, retain the customer.
Objections	”I never agreed to this”, “I’ll cancel my account”, “Get me your manager”.
Success criteria	Trainee acknowledges feelings within 15s, references policy verbatim, offers credit, doesn’t escalate.
Documents	`refund-policy-2026.pdf`, `customer-retention-playbook.md`
Rubric	Built-in `empathy + resolution + compliance`
Difficulty	Hard

Build a scenario

Open Training

Go to Training → Scenarios → New scenario.

Pick a persona

Either select an existing persona or click New persona to build one inline.

Write the opening and goal

Keep the opening to one or two sentences — that’s all the customer would actually say first. The goal is for you to remember why this scenario exists; the trainee never sees it.

List objections

3-5 specific pushbacks. The AI customer will weave them in as the conversation progresses.

Define success criteria

Either a checklist (acknowledged feelings, quoted policy, offered credit) or a free-text description. Used by the AI grader.

Attach documents

Drag in PDFs, Word docs, or Markdown. The trainee can open them mid-call (it’s open-book), and the AI customer might reference them too.

Pick a rubric

Use the built-in rubric to start. Customize per-scenario only if the situation needs it (e.g. compliance scripts that must be quoted verbatim).

Save and test

Run a self-attempt to make sure it feels right. Iterate.

Generate scenarios in bulk

Click Generate from description, paste a paragraph about your business and the kinds of calls you want to practice, and Omniflow drafts a batch of scenarios you can review.

Generated scenarios are a starting point, not the finished product. Always read each one — the AI is good at shape, less reliable on the specifics of your policies and product.

Open in Omniflow

If you want to…	Go to
Bundle scenarios into a path	Learning Paths
Run a practice call	Practice Calls
Customize the rubric	Scoring Rubric