> ## Documentation Index
> Fetch the complete documentation index at: https://docs.guhan.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Build your sequence

> Design the step-by-step flow that runs for every prospect.

Your sequence is what actually gets sent. Guhan gives you two views for building it — pick the one that fits your flow.

## Simple (stair) view

A vertical list of steps. Use for linear sequences with no branches.

```
Step 1  →  LinkedIn invite       [Edit]
Step 2  →  Wait 3 days           [Edit]
Step 3  →  LinkedIn DM            [Edit]
Step 4  →  Wait 4 days            [Edit]
Step 5  →  Email                  [Edit]
Step 6  →  Exit                    [—]
```

* Click **Edit** on any step → change the message brief, channel, timing
* Click **+** between steps → add a new step
* Reorder by drag

Simple view is enough for most workspaces. If your sequence has any branches, Guhan automatically switches you to Advanced.

## Advanced (canvas) view

A visual node graph. Use for:

* **Branches** based on prospect behavior
* **AI reply nodes** with multi-round approval flows
* **Complex conditions** (e.g., "only send email if prospect has an email address")

Drag nodes onto the canvas, click node handles to wire them, click nodes to configure.

## The node types

<AccordionGroup>
  <Accordion title="Start node">
    Where every prospect begins. One per sequence. Nothing to configure.
  </Accordion>

  <Accordion title="Send node — LinkedIn invite">
    Sends a LinkedIn connection invite.

    Configure:

    * **Note** (optional) — short custom note attached to the invite
    * **Template** — pick a Brand & Voice template as the starting point
    * **Sender** — which LinkedIn account sends this

    Max note length: 300 characters.
  </Accordion>

  <Accordion title="Send node — LinkedIn DM">
    Sends a direct message on LinkedIn. Requires prior connection.

    Configure:

    * **Brief** — what to say (Guhan drafts the actual message)
    * **Template** — starting point
    * **Sender**
    * **Fetch recent post** — opt-in; Guhan reads the prospect's latest post + uses it as a hook (1 credit)
  </Accordion>

  <Accordion title="Send node — Email">
    Sends an email.

    Configure:

    * **Subject brief** — Guhan writes the subject per prospect
    * **Body brief** — what to say
    * **Template** — starting point
    * **Sender** (Gmail / Outlook)
    * **Require approval** — hold send until you click Approve
  </Accordion>

  <Accordion title="Send node — WhatsApp">
    Sends a WhatsApp message. Requires prospect to have a revealed phone.

    Configure:

    * **Brief** — what to say
    * **Template**
    * **Sender** (WhatsApp)
  </Accordion>

  <Accordion title="Send node — Comment on LinkedIn post">
    Guhan drafts + posts a public comment on the prospect's most recent LinkedIn post. Great for warming a cold prospect before you DM them.

    Configure:

    * **Brief** — what kind of comment (question / observation / support)
    * **Fetch recent post** — required; Guhan needs a post to comment on
    * **Sender**
  </Accordion>

  <Accordion title="Send node — React to LinkedIn post">
    Reacts (like / celebrate / insightful) to the prospect's most recent post. Zero-cost warming step.

    Configure:

    * **Reaction type** — default: Like
    * **Sender**
  </Accordion>

  <Accordion title="Send node — View LinkedIn profile">
    Just views the prospect's profile (they'll see "who viewed your profile" if their privacy settings allow). Zero-cost warming step.
  </Accordion>

  <Accordion title="Wait node">
    Wait N hours or days before continuing.

    Configure:

    * **Duration** — 1 hour → 90 days
  </Accordion>

  <Accordion title="Wait for event node">
    Wait for a specific event (they replied, they accepted the invite, they opened your email) OR until a timeout expires.

    Configure:

    * **Event type** — invite accepted / LinkedIn replied / email opened / email clicked / email replied / any reply
    * **Timeout** — max wait (e.g., 7 days)
    * **On timeout** → which edge to take
  </Accordion>

  <Accordion title="Branch node">
    Split the flow based on rules about the prospect.

    Configure:

    * **Rule** — e.g., "has an email" / "in list X" / "posted in last 30 days" / "score > 70"
    * **Yes edge** → goes here if rule matches
    * **No edge** → goes here otherwise
  </Accordion>

  <Accordion title="AI reply node">
    When a reply comes in from a preceding wait\_for\_event, Guhan classifies + drafts a response.

    Configure:

    * **Send for approval** — you approve the draft before it sends (safer)
    * **Auto-send** — Guhan sends the draft immediately
    * **Max consecutive replies** — how many rounds Guhan handles before handing off to you
  </Accordion>

  <Accordion title="Handoff node">
    Pause the sequence, drop a task in your Inbox.

    Configure:

    * **Mode** — Notify / Approve / Decide / Take over / Schedule
    * **Subject** — task title
    * **Body** — instructions for whoever handles it
  </Accordion>

  <Accordion title="Exit node">
    End of the sequence. Prospect is marked done.
  </Accordion>
</AccordionGroup>

## A recipe: the standard 5-step flow

Great starting point for most B2B outbound:

```
Start
  ↓
LinkedIn invite (with brief note)
  ↓
Wait for event: invite accepted (timeout 14d → Exit)
  ↓
Wait 2 days
  ↓
LinkedIn DM (opt-in: fetch recent post)
  ↓
Wait for event: reply (timeout 5d → next)
  ↓
Email (short, references the DM they didn't reply to)
  ↓
Wait for event: reply (timeout 7d → next)
  ↓
Email follow-up (new value, no ask)
  ↓
Wait for event: reply (timeout 14d → Exit)
  ↓
Exit
```

Total sequence spans \~30 days. Typical reply rates:

* After invite: 40-60% acceptance
* After DM: 5-10% reply
* After first email: 2-4% additional reply
* After follow-up: 1-2% additional reply

Blended reply rate: \~10-15% across the full sequence.

## Editing a live sequence

You can edit anytime. When you click **Publish**:

* New prospects enrolled after publish use the new version
* Prospects already in-flight keep their old version (finish per what they started)

If you want to force in-flight prospects to move to the new version, use **Restart from step 1** on the sequence detail page (warning: may double-message them if they already got step 1).

## Related

<CardGroup cols={2}>
  <Card title="Message templates" icon="file-lines" href="/brand-and-voice/templates">
    Reusable briefs to start steps from.
  </Card>

  <Card title="Sequences" icon="diagram-project" href="/agents/outreach-agents/sequences">
    Concept overview.
  </Card>
</CardGroup>