> ## Documentation Index > Fetch the complete documentation index at: https://help.equip.co/llms.txt > Use this file to discover all available pages before exploring further. # Scenarios > Show candidates code, screen designs, emails or data during a conversational interview, and have the AI discuss it with them. A **Scenario** puts reference material in front of the candidate during a conversational interview. The candidate sees it in a panel on screen while the AI asks them about it, so you can test how someone works with real material, not just how they talk about it. Scenarios suit technical and non-technical roles alike. A developer can debug a code snippet, a designer can critique a screen, a support hire can reply to an upset customer, and a salesperson can handle a prospect's objection. > [!NOTE] > > Scenarios are rolling out gradually and may not yet appear in your Add item menu. ## Example Scenarios by Role Here is what a Scenario looks like for four different roles. In each image, the left side is what you enter and the right side is what the candidate sees. | Role | What the candidate sees | Markdown used | Answer format | | - | - | - | - | | **1. Software developer** | A short function with a bug | Code block | Written | | **2. Product designer** | A screen design to critique | Image | Spoken | | **3. Customer support** | An upset customer's email | Quote | Written | | **4. Sales** | A prospect's price objection | List | Spoken | ### 1. Software Developer Paste a short snippet with a bug in it and ask the candidate to find and fix it. The AI context tells the AI what the bug is, so it can probe and grade against it. ![Left, a Scenario item in the editor with a Candidate view titled Why does this crash containing a three-line Python average function, an AI context naming the ZeroDivisionError, and Answer format Written. Right, the candidate's screen with the rendered heading, code block and question, and a code answer box set to Python](https://mintcdn.com/equip-hq/z4aHVqMxa9CmzDpx/images/scenarios/code-review.png?fit=max\&auto=format\&n=z4aHVqMxa9CmzDpx\&q=85\&s=37e2c5a9c4b411046433bd7b61683eb9) Software developer: a Python function that crashes on an empty list ### 2. Product Designer Show a screen design as an image and ask what the candidate would change. The AI cannot see images, so the AI context describes what the screen shows and which problems to look for. ![Left, a Scenario item in the editor with a Candidate view linking to an image of a mobile sign-up form, an AI context describing the form's eight fields and vague error, and Answer format Spoken. Right, the candidate's screen showing the sign-up form image and the question, with no answer box because the answer is spoken](https://mintcdn.com/equip-hq/z4aHVqMxa9CmzDpx/images/scenarios/design-critique.png?fit=max\&auto=format\&n=z4aHVqMxa9CmzDpx\&q=85\&s=ed14f0bb0ccf0010e8e58311a704870a) Product designer: a mobile sign-up form most people abandon ### 3. Customer Support Quote a short customer email and ask the candidate to write a reply. The AI context sets out the right outcome and what a strong reply sounds like. ![Left, a Scenario item in the editor with a Candidate view quoting an email from Dana about being charged twice, an AI context describing a strong reply, and Answer format Written. Right, the candidate's screen showing the rendered email as a quote and a text answer box with a reply being typed](https://mintcdn.com/equip-hq/z4aHVqMxa9CmzDpx/images/scenarios/customer-reply.png?fit=max\&auto=format\&n=z4aHVqMxa9CmzDpx\&q=85\&s=5b55868c0caa84c5715690359ce2491c) Customer support: an email about a double charge ### 4. Sales Describe a prospect and their objection, then ask the candidate to respond as they would on a call. With a **Spoken** answer, you hear how they handle pushback in real time. ![Left, a Scenario item in the editor with a Candidate view titled Handle the price objection listing the prospect and their objection, an AI context describing a strong answer, and Answer format Spoken. Right, the candidate's screen showing the rendered heading, two bullet points and the question, with no answer box because the answer is spoken](https://mintcdn.com/equip-hq/rpdAx_7h6evpUWfs/images/scenarios/sales-objection.png?fit=max\&auto=format\&n=rpdAx_7h6evpUWfs\&q=85\&s=2794a56fc28a16050832f7cde2ffa0e5) Sales: a prospect says the product costs too much ## Adding a Scenario **1. Open the interview's settings** From the [AI Interviews dashboard](https://equip.co/recruiter/ai-interviews/dashboard/), open your conversational interview and go to the **Topics & questions** section. **2. Add a Scenario item** Click **Add item** and choose **Scenario**. **3. Write the candidate view** Enter the material the candidate should see, in Markdown. Start with a heading, add the material, and end with the question you want answered. **4. Add AI context (optional)** Describe anything the AI cannot read from the candidate view, such as an image, and the points a strong answer should cover. **5. Set difficulty, follow-ups and answer format** Pick a **Difficulty**, the number of **Follow-ups** (0 to 5), and an **Answer format** if the Workspace is on. During the interview, each Scenario appears as a numbered tab (Q1, Q2 and so on) above the candidate's answer area. The AI switches tabs as the conversation moves on. ## What You Can Put in the Candidate View The candidate view is written in **Markdown**, a simple way to format text with symbols. These elements are supported: | Element | How to write it | Use it for | | - | - | - | | **Heading** | `# Review this function` | The scenario's title | | **Bold** and *italic* | `**bold**`, `*italic*` | Key facts the candidate must not miss | | **List** | `- item` or `1. item` | Requirements, constraints, order details | | **Quote** | `> text` | Customer emails, messages, feedback | | **Table** | `\| Option \| Cost \|` rows | Options to compare, data, account details | | **Code block** | Code between two lines of three backticks | Code to review, logs, queries | | **Image** | `![Description](https://...)` | Screen designs, charts, diagrams | The candidate view and AI context can each hold up to **50,000 characters**. > [!WARNING] > > Images are added by URL only; you cannot upload them. Host the image somewhere the candidate's browser can open it without signing in, or it will not appear. ## Writing AI Context When you fill in **AI context**, the AI reads it instead of the candidate view. The AI also uses it when it evaluates the candidate's answer. The candidate never sees it. | Use AI context to | Example | | - | - | | Describe an image | *The image is a sign-up form with eight fields and a vague error message.* | | Set out what a strong answer covers | *Bug: it crashes when the list is empty. A good fix returns 0.* | | Share facts the candidate should work out | *The duplicate charge is refundable.* | > [!WARNING] > > Because the AI reads AI context in place of the candidate view, include a short summary of what the candidate sees. Otherwise the AI will not know what is on the candidate's screen. ## Choosing an Answer Format The **Answer format** decides how the candidate responds. It appears only when the [Workspace](https://help.equip.co/interview-workspace.md) is turned on for the interview. | Answer format | What the candidate does | Best for | | - | - | - | | **Spoken** | Talks it through with the AI; no answer box | Design critiques, case discussions | | **Written** | Types the answer in the Workspace, as text or code | Code fixes, email replies | | **Either** | Chooses whether to speak or write | Mixed tasks where both work | ## Related Resources * [Questions, Topics, and Scenarios](https://help.equip.co/interview-content-types.md) - How Scenarios compare with Questions and Topics * [The Workspace: Code and Written Answers](https://help.equip.co/interview-workspace.md) - Turn on written and code answers * [Creating a Conversational AI Interview](https://help.equip.co/create-conversational-interview.md) - Set up the interview your Scenarios go into * [AI Interview Reports](https://help.equip.co/interview-reports.md) - See how candidates answered each Scenario This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.