Skip to content

Creating Sequences

Step-by-step guide to creating and deploying outreach sequences in Sequence Studio.

Before you begin
  • ✅ HubSpot connected (Settings → General)
  • ✅ Contact lists created in HubSpot
  • ✅ Product positioning documented
  • ✅ Campaign goals defined
  1. Create Sequence

    Navigate to Sequence Studio and click + New

  2. Name Your Sequence

    Choose a descriptive name:

    • Include campaign type
    • Add date or version
    • Example: “INBOUND25 Outreach - Q1”
  3. Configure Generator

    Define how AI generates content

  4. Set Up Senders

    Configure who sends emails

  5. Define Steps

    Set cadence and touchpoints

  6. Configure Settings

    Set sources, limits, and policies

  7. Test

    Run with sample contacts

  8. Deploy

    Activate for production

  9. Monitor

    Track performance and adjust

Before configuration, document:

Campaign Goals:

  • What’s the objective? (meeting, demo, sale)
  • Who’s the target audience?
  • What’s the value proposition?

Content Strategy:

  • Tone (formal, casual, technical)
  • Length (brief, detailed)
  • Personalization level (high, medium, low)

Success Metrics:

  • Reply rate target
  • Meeting booking rate
  • Conversion goal

Default Instructions:

  1. Click + Add Prompt Part
  2. Select Default Instructions from library
  3. Choose preset (e.g., “Conference Default Instructions”)
  4. Review rules and adjust if needed

Custom Instructions:

  1. Add campaign-specific rules
  2. Define tone and style
  3. Set personalization boundaries
  4. Add compliance requirements

Content:

  1. Write product positioning:

    - Platform tagline
    - Key value propositions
    - Feature descriptions with benefits
    - Proof points and metrics
    - Case studies or testimonials
  2. Add campaign context:

    - Event details (if applicable)
    - Time-bound offers
    - Seasonal relevance
    - Competitive positioning

Toggle context available to AI:

HubSpot Sources:

  • ✅ Previous Communication (recommended)
  • ✅ Company Communication
  • ✅ Lifecycle Stages
  • ⬜ Deal Information (if relevant)

LinkedIn Sources:

  • ✅ Profile (for title/company verification)
  • ⬜ Posts (for high-value prospects only)
  • ✅ Company Profile
  • ⬜ Company Posts (if relevant)

External Sources:

  • ⬜ Company News (adds cost, use selectively)
  • ⬜ Financing Events (for investors/partners)
  • ⬜ Website Summary (for discovery calls)

LLM Model: Choose from available models (see LLM Models)

Recommendations:

  • gemini/gemini-2.5-pro — Best quality
  • openai/o3 — Good balance
  • openai/o4-mini — Fast and economical

Threading:

  • ✅ Send emails in one thread (recommended for follow-ups)

Auto-Enroll:

  • ⬜ Auto-enroll into CRM sequence (if using HubSpot sequences)
  • ⬜ Auto-enroll into Email Bison (if using)

Set the primary sender (usually you):

  • Name: John Smith
  • Email: john@company.com
  • Company: Acme Corp
  • Scheduling Link: calendly.com/john (optional)

For team sequences:

  1. Click + Add Sender
  2. Choose:
    • CRM User — Select from HubSpot users
    • Manual Entry — Enter details manually
  3. Configure selection logic:
    • Match by HubSpot owner
    • Match by territory
    • Round robin

Configure how contacts map to senders:

hubspot_owner_id: Maps to the contact’s assigned owner in HubSpot

territory: Custom field for regional assignment

account_executive: Custom field for AE assignment

Click Edit Steps to define cadence:

Example: 3-Email Sequence

Day 1: Initial Email
Day 3: Follow-up Email (same thread)
Day 7: Final Email (same thread)

Example: Multi-Channel Sequence

Day 1: Email
Day 3: LinkedIn Connection Request
Day 5: LinkedIn Message
Day 8: Email Follow-up
Day 12: Phone Call

Email:

  • Generates subject and body
  • HTML formatting
  • Plain text version
  • Threading support

LinkedIn Message:

  • Character limit enforcement
  • Connection request copy
  • Direct message copy

LinkedIn Connection:

  • Personalized connection request
  • Note text (300 char limit)

Phone Call:

  • Call script generation
  • Key talking points
  • Objection handling

Mark steps as optional:

  • ✅ Skip if previous step got reply
  • ✅ Skip if contact engaged elsewhere
  • ⬜ Required (always execute)

Active: Toggle to enable/disable sequence

Review Policy:

  • ⬜ Auto-confirm (immediate send)
  • ✅ Require Human Review (recommended initially)
  • Auto-confirm After: 10 minutes (if review enabled)

Pricing: Review credit cost per contact

Add HubSpot lists:

  1. Click + Add List
  2. Search and select list
  3. Set:
    • Priority: 1-10 (lower = higher priority)
    • Reschedule After: Days until retry
  4. Review list size (e.g., ~372 contacts)

Multiple Lists:

  • Contacts process by priority
  • Higher priority lists process first
  • Reschedule controls retry timing

Control speed and volume:

Enrollment Limits:

Max 1 contact / 5 seconds (system minimum)
Max 50 contacts / day (your limit)

Generation Limits:

Max 1 contact / 5 seconds (system minimum)
Max 50 contacts / day (your limit)

Weekend Blocking:

  • ✅ Do Not Allow Any on Weekends

Map generated content to HubSpot properties:

Per-Step Mapping:

Step 1:
- body → sellestial_email_1_body
- plain_text → sellestial_email_1_plain
- subject → sellestial_email_1_subject
Step 2:
- body → sellestial_email_2_body
...

Property Prefix:

  • Default: sellestial_
  • Custom: your_prefix_

Before deploying, check for contradictions:

  1. Click Prompt Consistency Check button in Generator tab
  2. AI analyzes all prompt sections
  3. Review identified issues:
    • Conflicting instructions
    • Contradictory rules
    • Incompatible requirements
  4. Fix issues in prompt
  5. Re-run check until clean
  6. Proceed to testing

Common Issues:

  • “Don’t include links” vs “Add scheduling link”
  • “Brief emails” vs “Include 3 case studies”
  • “Casual tone” vs “Formal business language”

Test with sample contacts:

  1. Go to Test tab
  2. Click Run button
  3. Review generated content:
    • Check personalization quality
    • Verify tone and length
    • Confirm fact accuracy
    • Test all steps
  4. Adjust Generator if needed
  5. Re-test until satisfied

What to Check:

  • ✅ All facts verifiable
  • ✅ No fabricated details
  • ✅ Appropriate personalization
  • ✅ Consistent tone
  • ✅ Proper formatting
  • ✅ Subject lines compelling
  • ✅ CTAs clear

Activate the sequence:

  1. Go to Deploy tab
  2. Review Generator Definition:
    • Current version hash
    • Deployed timestamp
    • Complete YAML spec
  3. Click Deploy button
  4. Confirm deployment
  5. Version activates
  6. Sequence status changes to Active

Deployment Banner:

If you see:

⚠️ Generator has changes that need to be compiled and deployed

This means edits in Generator tab aren’t live yet. Deploy to activate.

Track performance:

Monitor KPIs:

  • Total enrolled
  • Waiting for confirmation
  • Pending generation
  • Completed

Review individual results:

  1. Select contact
  2. View generated steps
  3. Read content
  4. Check for issues
  5. Approve or reject

Based on results:

  • Low reply rates → Adjust messaging
  • High rejection → Review personalization
  • Factual errors → Tighten prompt rules
  • Inconsistent quality → Add validation checks

Check:

  • Sequence is Active
  • Input lists have contacts
  • Processing limits not too restrictive
  • Lists aren’t already fully processed

Possible causes:

  • Too many data sources enabled
  • Large input lists
  • Complex prompts
  • Rate limits set too conservatively

Solutions:

  • Disable unnecessary data sources
  • Process smaller batches
  • Simplify prompts
  • Increase rate limits (if appropriate)

Issues:

  • Wrong tone
  • Incorrect facts
  • Poor personalization

Solutions:

  • Review Default Instructions
  • Add Custom Instructions
  • Run Consistency Check
  • Adjust prompt content
  • Test with more samples