# Setting Up Your AI Interview Agent

> This guide walks you through how to create, customize, test, and launch end-to-end hiring workflows with AI recruiting agents in HeyMilo. Everything here is designed to help recruiters move fast, stay consistent, and build high-performing interview flows with confidence.

## Before You Start

You'll need:

* A job description or job posting URL
* A clear understanding of what matters most in the role
* Access to your HeyMilo dashboard
* A general interview flow in mind (resume → SMS → voice/video, etc.)

{% hint style="success" %}
**Important:** The job description is used to generate questions, but it does NOT get passed directly into the AI agent’s underlying conversation prompt. If you want the AI to reference specifics from your JD during the interview, add those details manually inside the correct Context sections during setup.
{% endhint %}

{% stepper %}
{% step %} <mark style="color:blue;">**Step 1**</mark> Create a New Interviewer

Log in to your HeyMilo dashboard.

* Click the **"Create Interviewer"** button on the left-hand sidebar.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FRWdRNJ8EhGRETVnTzBgQ%2Fcreateinerviewrer%20(1).jpeg?alt=media&#x26;token=5f43c099-d9e1-4d58-b566-43913b6448d1" alt="" width="239"><figcaption></figcaption></figure>

This opens the New Interview Flow, where HeyMilo builds your custom interview experience.
{% endstep %}

{% step %} <mark style="color:blue;">**Step 2**</mark> Generate Your Interview Plan

You’ll see the “**Create an Interview Experience with a Job Description**” screen.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FqOkDD8NmLJP7DOSD2xfA%2Fwithjobdescirpiton%20(1).jpeg?alt=media&#x26;token=1d33757d-39f3-4e12-8502-db71226ae2c0" alt=""><figcaption></figcaption></figure>

You can:

{% tabs %}
{% tab title="Option A: Paste your job description" %}
Best for speed. HeyMilo parses it into core skills and responsibilities.
{% endtab %}

{% tab title="Option B: Enter a job posting URL" %}
HeyMilo extracts the text automatically.
{% endtab %}

{% tab title="Option C: Choose a quick-start template" %}
Good when you want inspiration or industry-standard questions.
{% endtab %}
{% endtabs %}

Click **Generate Interview Plan**.

**What HeyMilo does here:**

* Reads the JD
* Identifies responsibilities, skills, and expectations
* Creates a baseline interview plan
* Generates recommended stages and high-quality questions

{% hint style="success" %}
**Note:** *The JD is used for plan creation only — not passed into the live conversation prompt.*
{% endhint %}
{% endstep %}

{% step %} <mark style="color:blue;">**Step 3**</mark> Set Basic Interview Info

Customize how the interviewer appears to candidates:

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2Fj6OL3XApEgKHeerokH56%2Fcongigutesettibngs%20(1).jpeg?alt=media&#x26;token=27abe1a9-f796-4528-b4f7-c4a98a342cc9" alt=""><figcaption></figcaption></figure>

You’ll configure:

<details>

<summary>Interview URL Name</summary>

This becomes the candidate-facing link (e.g., `gointerview.co/yourcompany/full-stack-engineer`).

Best practices:

* Use lowercase + hyphens
* Keep it clear and searchable
* Make it specific to the role and location

Examples:

* `fce-full-stack-engineer`
* `sales-rep-nyc-2025`
* `warehouse-nightshift-denver`

Avoid:

* `agent1`
* `test`
* random strings

Changing this later changes the URL — so finalize before sharing.

</details>

<details>

<summary>Position Title</summary>

Auto-filled from your JD, but editable.

Click **Continue to Interview Workflow**.

**Why this matters:**

A clean URL and correct job title improve candidate trust and reduce drop-off.

</details>
{% endstep %}

{% step %} <mark style="color:blue;">**Step 4**</mark> Build Your Interview Flow

You’ll now choose which screening stages to include.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FxbAkJPmYdOWj4P33Dln0%2Fcdsacdsasss%20(1).jpeg?alt=media&#x26;token=4b9a57fd-1813-44ef-bf07-612870de75db" alt=""><figcaption></figcaption></figure>

Available agents:

* **Voice Interview (Sarah)**
* **Resume Screening (Roscoe)**
* **SMS Screening (Tyler)**
* **Form Screening (Fred)**

Drag stages into your workflow.

<details>

<summary><strong>Example workflows</strong></summary>

* Resume → Voice Interview
* SMS → Phone
* Resume → SMS → Voice
* Form → Resume → Voice

</details>

<details>

<summary><strong>Cost Projector (Live Cost Breakdown)</strong></summary>

As you add and reorder stages, HeyMilo shows a live cost projection for your interview flow so there are no surprises later.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FvIqNm1RbH6I7tdxeRYk4%2Fbfdsbfdvfds.jpeg?alt=media&#x26;token=ffe48967-7588-4b5e-a46a-8e5b1ecdf8ac" alt=""><figcaption></figcaption></figure>

The cost projector updates in real time and shows:

* **Cost per stage** (SMS, Resume, Voice, Form)
* **Estimated total cost** based on your workflow
* **Candidate drop-off assumptions** between stages
* **How many candidates are expected to reach each step**

This helps you quickly answer questions like:

* *What happens to cost if I add a voice interview?*
* *Should SMS come before or after resume screening?*
* *How much does this flow cost at scale?*

You can drag stages around and immediately see how cost changes before moving forward. Optionally use this to decide on a budget in the later steps.

</details>

<details>

<summary><strong>View Pricebook (Cost Preview)</strong></summary>

Before moving on, click View Pricebook to see projected costs for your interview flow. Pricing updates automatically based on the agents and stages you add, helping you fine-tune your workflow before activation.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FbP0o6fer75xzyUYFLOYh%2FUntitledgrsgre.jpeg?alt=media&#x26;token=a53155c2-547e-451a-bdfa-567707acde0b" alt=""><figcaption></figcaption></figure>

</details>

<details>

<summary><strong>Essential Workflow Settings</strong></summary>

These settings apply across your entire interview experience and control how candidates see, access, and complete the interview.

**Agent Identity (Required)**

Choose how your AI agents identify themselves to candidates.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2F55cs3jOnRP3PHU30Y6W6%2Fvfdsbfebfdsb.jpeg?alt=media&#x26;token=5b463785-beb6-47f8-a5a1-3450600bed51" alt=""><figcaption></figcaption></figure>

* **Single Identity (Seamless Experience):** One agent name is used across all interview stages. Best for a consistent, human experience.
* **Multiple Identities (Handoff Experience):** Each interview stage can have its own agent name. Configure names directly within each workflow stage. Agent names appear on each stage card.

**Default Interviewer Language**

Select the primary language your AI interviewer will use.

* Supports 15+ languages and dialects
* **Multilingual Auto-Detection:**

  The interview starts in the default language, then adapts automatically if the candidate responds in another language.

**Timeline & Completion (Optional)**

* **Interview Deadline**

  Set a hard cutoff date. Candidates cannot start or continue after this date.
* **Days to Complete**

  How long candidates have to finish once invited. Timely nudges, and a final reminder is sent before expiration.
* **Interview Retake Policy**

  Allow or block retakes if candidates don’t pass.

**Communication & Branding (Optional)**

* **Email Template Group**

  Customize candidate emails for branding or compliance.
* **Phone Number**

  Used for SMS and voice interviews (if enabled).
* **SMS Invitations**

  Send interview links and reminders via text when a phone number is available.
* **Email Report to Candidate**

  When enabled, candidates automatically receive an email with a link to view and download their interview report (PDF), which includes the same scoring and analysis your team sees.

**Links & Redirects (Optional)**

* **Scheduling Link**

  Shows a “Schedule Interview” button once candidates meet your criteria.
* **Redirect Link**

  Sends candidates to a specific page after they complete all interview stages.

**Advanced Settings (Optional)**

Used to give the AI deeper context and improve interview quality:

* Company Information
* Role Details
* Hiring Process Overview
* Interview Style & Approach
* Sender Email (send from your own domain)

</details>

Click **Preview** anytime to see the candidate-facing design.

{% hint style="success" %}
**Why this matters:** This step determines the *order* and *logic* of how candidates move through your process — the backbone of your AI recruiting workflow.
{% endhint %}
{% endstep %}

{% step %} <mark style="color:blue;">**Step 5**</mark> Configure Your Desired Screening Stages

<mark style="color:$primary;">**Voice / Video Interview Setup**</mark>

This is your main conversational interview experience.

Settings Tab:

This controls **how the interview runs and behaves**.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2Fn4rQ6IB7XSMQvOcJhYn1%2Fcdacdassss.jpeg?alt=media&#x26;token=2e023431-947f-4cdf-b160-1d46165faabc" alt=""><figcaption></figcaption></figure>

{% hint style="success" %}
**Accent Options:** HeyMilo offers voices with different accents including American, British, Indian, Australian, and more. Select your preferred voice from the dropdown to match your candidate audience. Watch the video below to learn how to select accents!

Voices are from LMNT and Cartesia. To browse and add more, go to AI (sidebar) → Languages & Voices.
{% endhint %}

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2Fwkt18TwrS8nROUI5gfIw%2FRachel%20Valentino_s%20Video%20-%20Feb%209%2C%202026.mp4?alt=media&token=84860e6f-aa3a-4fca-9cad-1314f967329d>" %}

<details>

<summary><strong>Identity</strong></summary>

* **Agent Name:** pulled from Essential Settings (e.g. *Sarah*)
* **Voice:** selected AI voice
* Uses a unified identity across all stages unless overridden

</details>

<details>

<summary><strong>Modality</strong></summary>

Choose how candidates interview:

* **Voice**
* **Video**
* **Phone Calls** (if numbers are provisioned)

You can enable multiple options if needed.

</details>

<details>

<summary><strong>Experience</strong></summary>

* **Show Transcript:** optionally display a live transcript to candidates

Useful for accessibility and transparency. Transcripts are always saved internally.

* **Enable Cheat Detection:** toggle detection on or off

</details>

<details>

<summary><strong>Cheat Detection Controls</strong></summary>

* **Detection Threshold (0–100):** controls strictness
* **Detection Types**:
  * Facial behaviour
  * Multiple people
  * Phone detection
  * AI/scripted answers
  * Unusual delays
  * Tab switching

</details>

<details>

<summary><strong>AI Model</strong></summary>

* **Standard Model:** clear, predictable
* **Enhanced Model:** smoother, interruption-aware
* **Advanced Model**\*:\* most humanlike

</details>

<details>

<summary><strong>Advanced (Visual + Behavior)</strong></summary>

**Where:** Voice/Video interviewer → scroll to the **Advanced** section.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FhHYihe7RJAI1RQrI3nmZ%2FScreenshot%202026-04-16%20at%209.59.33%E2%80%AFAM.png?alt=media&#x26;token=12df849e-9a93-404a-a5e6-a2c4e30f4f06" alt=""><figcaption></figcaption></figure>

Start with the **defaults** and adjust one field at a time.

**What you can customize**

* **Backdrop:** background for voice interviews
* **Avatar:** interviewer visual
* **Timing before AI speaks / after interruption** — Replaces the old **Low / Medium / High** preset with exact values you set.
* **Max reminders** — Maximum number of times the AI can re-prompt a candidate who has gone quiet.
* **Silent seconds before prompt** — How long the AI waits (in seconds) before nudging the candidate.
* **Warn before ending** — Give the candidate a **warning** before the interview ends. Set how many seconds before the end to warn.
* **End interview after N seconds** — How long of silence / inactivity triggers the interview to end.
* **Warning attempts** — How many warnings the AI gives before it actually ends the interview.
* **Minimum speech clarity** — Answers scoring below this threshold are **flagged for review** (informational; helps you spot low-quality audio or unclear responses).

**Built-in validation**

Some values **must be consistent with each other**. For example, **Warn before ending** must be **shorter** than the **End interview after** time. If the values do not make sense together, the product will flag it and prevent you from saving until you fix it.

**How to tune it**

1. Start with the **default values**.
2. Change **one field at a time**.
3. Run **Test Now** as a candidate and listen for pacing.
4. Adjust and re-test until the flow feels natural for your role.

{% hint style="info" %}
**Tip:** Treat this as fine-tuning. For most teams the defaults are a good starting point, and small changes (for example trimming a few seconds off silence tolerance) go a long way.
{% endhint %}

</details>

Conversation Tab:

This controls **what the AI asks, how it evaluates, and how it speaks**.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FM2eKxr6uLcJhkkqaubnJ%2Ffrfrfre.jpeg?alt=media&#x26;token=bac97986-edc5-48b9-993c-cffd9fe8d6e8" alt=""><figcaption></figcaption></figure>

<details>

<summary><strong>Structure</strong></summary>

* **Number of Questions:** 1–10 (5 recommended)
* Sets how many questions will be auto-generated in the next step

</details>

<details>

<summary><strong>Qualification Requirements</strong></summary>

* **Enforce Objective Evaluation**
* Set a **Minimum Score (0–100%)** candidates must reach to pass

</details>

<details>

<summary><strong>Messages</strong></summary>

Optionally customize how the interviewer communicates:

* **Introduction Message**
* **Welcome Back Message**
* **Web Interview Closing Message**

You can load defaults and use variables like:

`{{name}}`, `{{interviewer_name}}`, `{{company_name}}`, `{{job_title}}`

</details>

<details>

<summary><strong>Context</strong></summary>

* **Enable Context:** shared across all stages (override per stage if needed)
* **Populate from JD** (optional starting point)

Use Context to add:

* core traits you’re screening for
* required tools or experience
* what “good” looks like for this role
* hiring process details

{% hint style="warning" %}
Reminder: The full job description is **not** passed into the live conversation prompt unless you include key details here.
{% endhint %}

</details>

<details>

<summary><strong>Conversation Style</strong></summary>

Choose how the AI speaks:

* **Formal & Direct:** structured, compliance-friendly
* **Balanced & Professional:** most commonly used
* **Relaxed & Natural:** conversational and warm

</details>

<mark style="color:$primary;">**Resume Screening Setup**</mark>

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FU36YQilgKEsySRp9z3TX%2Fcdsacdsacsss%20(1).jpeg?alt=media&#x26;token=2be2b7a3-439a-46b6-bf03-a5e76b700a9f" alt=""><figcaption></figcaption></figure>

Configure:

* Require resume upload (yes/no)
* Number of required qualifications (1-8)
* Pass/fail dealbreakers
* 1–10 scoring criteria
* Minimum score to advance
* Resume relevance signals

Perfect for filtering large applicant pools.

<mark style="color:$primary;">**SMS Screening Setup**</mark>

SMS Screening is optimized for fast, high-volume qualification and candidate engagement.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FwrWGemilT7SB4zJI6hk0%2Fbfdsvfdsvdsa.jpeg?alt=media&#x26;token=824919cb-c6b5-428d-9d08-bbd43f75913c" alt=""><figcaption></figcaption></figure>

Quick Setup & Configuration Tab:

<details>

<summary><strong>Prompt Version</strong></summary>

* **V1 (legacy)**
* **V3 (latest)**

  Includes custom messages, multilingual support, and advanced nudges

</details>

<details>

<summary><strong>Phone Number</strong></summary>

Pulled from Interview Stages or Essential Settings

</details>

<details>

<summary><strong>Multilingual</strong></summary>

Enabled from Essential Settings (if configured)

</details>

<details>

<summary><strong>Agent Name</strong></summary>

Uses unified name from Essential Settings

</details>

<details>

<summary><strong>Structure</strong></summary>

* **Number of Questions:** 1, 3, or 6 (3 recommended)
* **Enforce Required Qualifications:** auto-disqualify if criteria aren’t met

</details>

Customize Messages Tab:

Control exactly how candidates experience SMS conversations.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FYYuz4renAYEsfmZdjnzr%2Fhtrehtrerr.jpeg?alt=media&#x26;token=62d26ade-5e2c-4f02-a9a5-2c695cbbfa11" alt=""><figcaption></figcaption></figure>

<details>

<summary><strong>Intro Message</strong></summary>

Sent when the conversation starts

Uses variables like:

`{{name}}`, `{{interviewer_name}}`, `{{company_name}}`, `{{job_title}}`

</details>

<details>

<summary><strong>Disqualification Message</strong></summary>

Sent when a candidate fails knockout criteria

</details>

<details>

<summary><strong>Follow-up Message</strong></summary>

Used to nudge candidates who haven’t responded

</details>

<details>

<summary><strong>Call Invitation Message</strong></summary>

Sent to qualified candidates to move them into a phone or voice interview

You can:

* Load defaults
* Customize copy
* Let AI generate messages automatically (toggle on/off)

</details>

<details>

<summary><strong>Context</strong></summary>

* **Enable Context:** shared across stages
* **Populate from JD** (optional)

Give the agent guidance on interview style & approach, company information, role details, hiring process overview in this section.

</details>

<details>

<summary><strong>Nudge &#x26; Timing Settings</strong></summary>

**Cooldown Period**

* Hours to wait before nudging a candidate (e.g. 24 hours)

**Maximum Nudges**

* Max follow-ups per candidate (e.g. 3)

**Blackout Period**

* Time window when messages should not be sent

  Helps protect candidate experience and compliance across time zones

</details>

<mark style="color:$primary;">**Form Screening Setup**</mark>

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2F6Z79LnWGopPMLRo3I8Ki%2Fcdsaaaaa%20(1).jpeg?alt=media&#x26;token=5546c2a0-8b8c-4166-a28b-df42ca01fbd1" alt=""><figcaption></figcaption></figure>

Use structured form questions to collect:

* Experience
* Certifications
* Availability
* Files (e.g., licenses)
* Numeric or date values

Dealbreakers can automatically move only qualified candidates forward.
{% endstep %}

{% step %} <mark style="color:blue;">**Step 6**</mark> Preview The Flow You Built

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FJ6xZzFIHJOhgj2ErE4ta%2Fcdsaaaaaqq%20(1).jpeg?alt=media&#x26;token=71c90685-ba98-48ae-8830-c1b37f880b92" alt=""><figcaption></figcaption></figure>

Use **Preview** to confirm:

* Flow order
* Messages
* Question relevance
* Branding

Previews reduce errors and ensure a polished experience.
{% endstep %}

{% step %} <mark style="color:blue;">**Step 7**</mark> Set Up Your Questions & Scoring (For Every Interview Stage)

This is one of the most important parts of setting up your interview experience.

Every agent type — **Voice/Video, Resume, SMS, and Form Screening** — includes a *Questions & Scoring* section.

This is where you define **what good looks like** and how the AI evaluates candidates.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2Fur6p6mvvavF9Omb9QZfV%2Ffewfwwwww%20(1).jpeg?alt=media&#x26;token=86628bb2-cf9e-4e54-a6d9-78cf79f62ffa" alt=""><figcaption></figcaption></figure>

**⭐ What You Can Do in Questions & Scoring (All Agents)**

Across every screening stage, you can:

<details>

<summary>Edit or Replace Any Auto-Generated Content</summary>

HeyMilo always generates a structured baseline for you:

* 5 core interview questions (Voice/Video)
* Recommended resume criteria
* Suggested SMS dealbreakers
* Sample form intake questions

You can keep, edit, delete, or add your own.

</details>

<details>

<summary>Add New Questions or Criteria</summary>

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FF4y1yttnUIefckhleCZD%2Ffdawwwww%20(1).jpeg?alt=media&#x26;token=9430dae1-9d18-4b42-9056-c14527090b9a" alt=""><figcaption></figcaption></figure>

Each question/criterion supports:

* **Scored** (1–5 scoring)
* **Structured** (pass/fail or expected-answer logic)
* **Weighted scoring** (to prioritize certain questions)
* **Follow-up ranges** (how many follow-up questions the AI should ask)

This lets you tailor evaluations to the job’s true requirements.

</details>

<details>

<summary>Customize Evaluation Criteria</summary>

For scored questions, define:

* What a strong answer looks like
* What a weak answer looks like
* Additional instructions for the AI interviewer

For structured questions, define:

* The correct answer
* Accepted formats (YES/NO, numeric, multiple choice)
* Minimum values or required selections

This ensures consistent scoring across candidates.

</details>

<details>

<summary>Use Required Qualifications (Auto-Disqualifiers)</summary>

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2F4DQdlPZAqLmlSmDGanHT%2Fcdsacdsaaaa%20(1).jpeg?alt=media&#x26;token=46f96126-e307-4024-803c-3fad5274332d" alt=""><figcaption></figcaption></figure>

You can mark certain items as **dealbreakers**, such as:

* Work authorization
* Years of experience
* Certifications
* Schedule/shift availability

If a candidate fails these, HeyMilo auto-disqualifies them and stops further stages (if enabled).

Works across:

* Resume Screening
* SMS Screening
* Form Screening
* Voice/Video (via structured questions)

</details>

<details>

<summary>Add Tags (Optional)</summary>

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2Ff6StP7SD3gIG6WhbkWLn%2Fdfsafdsss%20(1).jpeg?alt=media&#x26;token=06708697-8700-4e99-a948-2bcf4beb6089" alt=""><figcaption></figcaption></figure>

Tags listen in the background and help you surface specific information, to help you flag important candidate details such as:

* Location
* Work authorization
* Certification
* Salary expectations
* Shift availability

Tags appear in the candidate’s report but **do not affect scoring**.

</details>

<details>

<summary>Test Your Questions Before Going Live</summary>

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2F9W6UqCmuVEcMWNICE6cs%2Fvcvbcbcvbhh%20(1).jpeg?alt=media&#x26;token=3a771e61-46f2-4ec6-ba74-6246aa2781cf" alt=""><figcaption></figcaption></figure>

Click **Test Now** at the bottom of any Questions & Scoring section.

This allows you to experience:

* how the AI asks your questions
* how follow-ups behave
* how scoring logic triggers
* how candidates see the flow

If something feels unclear, you can revise questions immediately.

</details>
{% endstep %}

{% step %} <mark style="color:blue;">**Step 8**</mark> Configure Email Alerts & Optional Budget

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FhbUXfOEQfmYSI8FcA8jG%2Fcdsacgggghhh%20(1).jpeg?alt=media&#x26;token=684ad730-44cd-409f-aa63-617436604569" alt=""><figcaption></figcaption></figure>

**Email Subscribers**

* Team members receive a notification when interviews are completed.

**Budget Allocation**

* Useful for predictable cost control.

You can set:

* Total budget
* Daily budget
* Maximum interviews allowed

Protects cost during large hiring pushes.
{% endstep %}

{% step %} <mark style="color:blue;">**Step 9**</mark> Activate & Share

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FdztWbylh1pJW2K3sNZpK%2Ffewafewwwww%20(1).jpeg?alt=media&#x26;token=d0b980cb-610f-4983-8938-b01fc1ac9d65" alt=""><figcaption></figcaption></figure>

After you click **Create Interview & Generate Questions,** you’ll now see all sharing and integration options in one place.

<details>

<summary><strong>Interview Link</strong></summary>

Your main candidate-facing URL to include in job ads, SMS, email outreach, or internal referrals.

</details>

<details>

<summary><strong>Printable QR Poster</strong></summary>

Generates a branded PDF with a QR code for walk-ins, hiring events, and on-site recruiting.

</details>

<details>

<summary><strong>ATS Integration</strong></summary>

Connect to Bullhorn, Avionté, JobDiva, Greenhouse, and more so candidates and completed interviews sync automatically.

</details>

<details>

<summary><strong>Webhook Ingestion</strong></summary>

A webhook URL is provided to automatically send candidates from internal systems, job boards, or automations (Zapier, Make, custom scripts).

</details>

<details>

<summary><strong>Indeed Email Ingestion</strong></summary>

A unique email address is generated for each interview. Add it to your Indeed job posting to auto-ingest all applicants into your pipeline and trigger interviews instantly.

</details>

<details>

<summary><strong>Ingestion Gateway</strong></summary>

A bulk import portal for uploading large candidate lists or migrating applicants from external sources (CSV upload, API-based ingestion, tagging options).

</details>

These ingestion methods ensure all candidates flow into your interview experience automatically, keeping your pipeline clean, fast, and fully automated.
{% endstep %}
{% endstepper %}

## Managing Agent Integrations After Creation

Once your agent is live, the **Integrations** tab provides all the ways candidates can enter this specific interview. It's organized into sections:

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FCKStVpH52H3T4p9uOgYh%2FcdsCDACSACA.jpeg?alt=media&#x26;token=26b42b17-2b8e-4332-9dae-c124ce7a9790" alt=""><figcaption></figcaption></figure>

<details>

<summary><strong>ATS Integration</strong></summary>

Connect this agent to your ATS so candidates and results sync automatically. If no ATS is configured, you'll see a prompt to set one up.

</details>

<details>

<summary><strong>Indeed</strong></summary>

Import candidates coming directly from Indeed job board traffic into this agent.

</details>

<details>

<summary><strong>Inbound</strong></summary>

Ways candidates can enter the interview:

* **Share Link:** the candidate-facing interview URL you can copy and share
* **Add Manually:** enter candidate details directly
* **Ingestion Gateway:** bulk upload resumes or candidate files (PDF, CSV, ZIP)
* **Ingestion API:** add candidates programmatically via API
* **Apply Email:** a unique email address for this agent; candidates who email it are auto-added
* **Backfill Candidates:** re-import or backfill candidates from other sources

</details>

<details>

<summary><strong>Outbound</strong></summary>

Outbound communication and engagement options for this agent.

</details>

<details>

<summary><strong>Webhooks</strong></summary>

Configure webhook endpoints for this agent to push events (interview started, completed, report available) to external systems.

</details>

## What Happens Next?

You get a full candidate profile including:

1. **Automatic Analysis**: AI evaluates all responses
2. **Scorecard Generation**: Detailed candidate profiles created
3. **Dashboard Updates**: View candidate reports and updates in HeyMilo or your connected ATS
4. **Team Collaboration**: Share read-only reports with hiring teams or clients
5. **Export Options**: Download PDFs or share links for easy access and review
6. **Notifications**: Alerts when candidates complete interviews

## Edit or Clone a Workflow After Candidates Have Started

Once candidates have completed an interview, HeyMilo **locks questions and scoring** to protect consistency. But you can still edit other agent settings anytime.

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FQMe8f0lxTBkE86VCWqzk%2FRachel%20Valentino_s%20Video%20-%20Feb%2011%2C%202026.mp4?alt=media&token=af5f0388-6dc3-4c82-8b41-e00b17b098e0>" %}

#### Where to Find It

Click into your agent. The **Settings** panel appears on the right with expandable sections.

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FiFRll19v3ZQ0bMqWvnRT%2Fbbfdsbtg.jpeg?alt=media&#x26;token=9b8e2bf3-caaf-4864-9ce3-dc08eb9b6b78" alt=""><figcaption></figcaption></figure>

<details>

<summary><strong>Agent Overview</strong></summary>

At the top of the Settings panel, you'll see the **Agent Overview** showing:

* **Agent Mascot:** the AI persona assigned to this agent
* **Status badge:** Live, Expired, or Archived
* **Created / Updated timestamps:** when the agent was created and last modified
* **Time remaining:** days until the agent expires (e.g., "5 days remaining")
* **Test Now** button: run through the interview yourself to verify the experience

</details>

#### What You Can Edit After Go-Live

<details>

<summary><strong>Root Level Configuration</strong></summary>

* Interviewer Name

</details>

<details>

<summary><strong>Advanced</strong></summary>

* Days to Complete
* Scheduling URL
* Redirect URL

</details>

<details>

<summary><strong>Additional Context</strong></summary>

* **Company Overview** (max 800 characters)
* **Job Overview** (max 800 characters)
* **Interview Process** (max 800 characters)
* **Instructions** (max 2000 characters)

</details>

<details>

<summary><strong>Per Workflow Step (e.g., Video Interview, Resume)</strong></summary>

Expand any workflow step under **Workflow Steps** to edit step-specific settings:

* **Interviewer Name** (step override)
* **Voice** (select voice)
* **Introduction Message**, **Welcome Back Message**, **Closing Message**
* **Company Overview**, **Job Overview**, **Interview Process**, **Instructions** (step-level overrides with Usage Mode: Reference Only, Summarized Delivery, or Exact Delivery)
* **Tags** (edit or delete existing tags)

</details>

<details>

<summary><strong>Communication Channels</strong></summary>

* Email Template Group

</details>

<details>

<summary><strong>Customizations</strong></summary>

* Design Template Group

</details>

<details>

<summary><strong>Advanced Settings</strong></summary>

* Email Report to Candidate (toggle on/off)

</details>

Click the **pencil icon** or **"Click to edit"** on any field, make your changes, and save. Changes apply to future candidates only.

#### What You Cannot Edit

**Questions and scoring criteria** are locked once candidates complete interviews. This prevents mixed scoring logic and keeps evaluations consistent.

#### Cloning an Agent: When You Need to Change Questions or Scoring

If you need to update questions, scoring, or workflow stages:

1. Go to your agent
2. Click **Actions**
3. Select **Duplicate Interviewer**

<figure><img src="https://110840537-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FbKCgmPgkC7X1VTl62ocK%2Fuploads%2FLdC4INK9Y5D60VFljAri%2Fbfsdbfdsbfd.jpeg?alt=media&#x26;token=d3244f60-14ce-4c94-bc3a-ea0afafafb5d" alt="" width="247"><figcaption></figcaption></figure>

4. Edit questions, scoring, or stages in the new version
5. Share or connect the new interview going forward

Your original interview remains intact for earlier candidates. The new version applies only to future candidates.

Don't forget to unlink the previous agent and reconnect the new interviewer to your ATS if you were using

{% hint style="success" %}
Note: If you clone an agent and change the job description, it won't auto-update the interview. You'll need to regenerate or adjust content manually. Cloning preserves the original questions and messaging.
{% endhint %}

## Common First-Time Tips

<details>

<summary>✅ Start Simple</summary>

* Begin with 5-7 core questions
* Add complexity as you get comfortable
* Focus on your most important requirements

</details>

<details>

<summary>✅ Test Thoroughly</summary>

* Complete the interview yourself first
* Ask a colleague to test it
* Refine based on the experience

</details>

<details>

<summary>✅ Monitor Early Results</summary>

* Review your first 5-10 candidates carefully
* Create a new agent and adjust scoring criteria if needed
* Fine-tune questions based on responses if needed

</details>

<details>

<summary>✅ Iterate and Improve</summary>

* Your first agent won’t be perfect. Once candidates complete your workflow, you’ll need to **duplicate the interview** to make changes. This preserves scoring consistency while letting you iterate safely.
* Use candidate feedback to improve
* Continuously refine your approach

</details>

## Common Configuration Mistakes to Avoid

<details>

<summary>❌ Too Many Questions</summary>

* Limit to 5-7 core questions initially
* Long interviews reduce completion rates
* Focus on most critical qualifications

</details>

<details>

<summary>❌ Unclear Instructions</summary>

* Provide clear guidance to candidates
* Explain the process and expectations
* Include technical requirements

</details>

<details>

<summary>❌ Overly Restrictive Knockout Criteria</summary>

* Don't eliminate good candidates unnecessarily
* Consider "preferred" vs "required" qualifications
* Allow for equivalent experience

</details>

<details>

<summary>❌ Inconsistent Scoring</summary>

* Ensure scoring criteria align with job requirements
* Weight questions appropriately
* Test scoring with sample responses

</details>

{% hint style="warning" %}
**Remember**: You can only modify your agent settings after creation when candidates haven't interacted with it yet. Alternatively, you can clone your agent to keep the same settings. It's important to start with a good foundation and refine based on real candidate interactions.
{% endhint %}


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.admin.heymilo.ai/md2Ig4yfyjUbBviADESH/creating-your-interview-agents/setup.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.