Home› Blog› AI & Automation› Customer Journey Automation: A Practical Guide
AI & Automation

Customer Journey Automation: A Practical Guide

A B2B SaaS company’s real customer journey, written out step by step, exposes exactly where follow-up depends on someone remembering to do it. Here’s how to map yours and turn each gap into something that runs on its own.

Vance Moore
Vance Moore
August 7, 2026
12 min read
2,712 words
Customer Journey Automation: A Practical Guide

Customer journey automation connects the touchpoints a prospect moves through, an ad click, a landing page visit, an email, a demo booked, so each step triggers the next one automatically based on what the prospect actually does. For a B2B SaaS company, it means a trial user who goes quiet gets a relevant follow-up instead of falling out of the pipeline unnoticed.

This guide shows how to build it in five steps: map one real journey, find the gaps in your own data, turn each gap into a trigger, test the rules on sample events, and review them monthly. It includes a query and a rules script that run as shown, a trigger template you can download, and a short walkthrough video. Skill level: comfortable with a spreadsheet and a basic script. Time: about half a day for a first journey.

Why do most B2B SaaS teams have an incomplete journey?

Most B2B SaaS teams can describe their customer journey in general terms: someone finds us, they reach out, we follow up, they book. What is usually missing is what happens at each specific gap, the one to three days between a demo request and the first callback, or the two to three weeks between a proposal and a decision.

Those gaps are where leads are lost. Not because the business did anything wrong, but because following up consistently, at the right time, with the right message, is hard to do by hand across dozens or hundreds of leads a month.

Diagram of a B2B SaaS journey from ad click to decision with three gaps marked: after the form, before the callback, and after the proposal
The same journey with the three gaps marked. Illustrative journey; every gold marker depends on a person remembering to act.

What do you need before you start?

Item What you need Why
Data Timestamps for form submissions, first callback, proposal sent, and replies Gaps are measured from these events, so export them from your CRM or email tool
Tools A spreadsheet, Python 3, and your CRM or automation tool The scripts below use only the Python standard library
Time About half a day for one journey Most of it is mapping and agreeing the thresholds
Result A trigger map, tested rules, and a monthly review habit Plus a downloadable template and checklist
Silent walkthrough (about 50 seconds) of the five steps with the script output. The same text is in the sections below.

Step 1: Map a real journey before automating it

Write out an actual journey a real customer went through recently, not a generic funnel. For a B2B SaaS company, that might look like this:

  1. Sees a LinkedIn ad and clicks through to a landing page
  2. Signs up for a free trial or requests a demo
  3. Gets a confirmation email (sometimes)
  4. Waits for a sales follow-up, which may take one to three days
  5. Has a demo call scheduled
  6. Receives a pricing proposal
  7. Goes quiet for two to three weeks while evaluating internally
  8. Either converts or the lead goes cold with no further contact

Written out this way, the gaps are obvious: the inconsistent confirmation, the one-to-three day callback delay, and the total silence after the proposal is delivered. Each is a specific point where automation can close the gap, not with a vague “improve follow-up” initiative, but with a defined trigger and action. Capture each stage in a table so nothing stays implicit.

Stage Touchpoint Owner today What the prospect sees Time to next step
Form submitted Landing page form Marketing Thank-you page, sometimes an email Minutes to never
First callback Phone or email from sales Sales rep Nothing until contacted 1 to 3 days
Demo Video call Sales rep Calendar invite Days
Proposal Emailed document Sales rep The document 2 to 3 weeks of silence

Common mistake: mapping the journey you wish you had. Use a real lead and the timestamps in your CRM.

Step 2: Find the gaps in your own data

Before building anything, measure how long leads actually wait. Export the created time and the first-touch time for recent leads, then list every lead whose wait passed your threshold. The query below runs on a scratch SQLite database with six illustrative leads, using a 12-hour threshold and a fixed “now” so the output repeats.

find_gaps.py (swap the scratch table for your own export)

import sqlite3

NOW = "2026-10-12 09:00"      # fixed so the output is repeatable
THRESHOLD_HOURS = 12         # flag leads that waited longer than this for a first touch

db = sqlite3.connect(":memory:")
db.executescript("""
CREATE TABLE leads (id TEXT, created_at TEXT, first_touch_at TEXT);
INSERT INTO leads VALUES
 ('L1', '2026-10-06 09:10', '2026-10-06 15:00'),
 ('L2', '2026-10-09 14:30', NULL),
 ('L3', '2026-10-01 11:00', '2026-10-02 09:00'),
 ('L4', '2026-09-20 10:00', '2026-09-21 10:00'),
 ('L5', '2026-10-11 16:00', NULL),
 ('L6', '2026-10-12 08:30', NULL);
""")

query = """
SELECT id, created_at, COALESCE(first_touch_at, 'none yet') AS first_touch, hours_waiting
FROM (
  SELECT id, created_at, first_touch_at,
         ROUND((julianday(COALESCE(first_touch_at, ?)) - julianday(created_at)) * 24, 1) AS hours_waiting
  FROM leads
)
WHERE hours_waiting > ?
ORDER BY hours_waiting DESC;
"""
for row in db.execute(query, (NOW, THRESHOLD_HOURS)):
    print(row)

Output (illustrative sample data, columns: lead, created, first touch, hours waiting):

('L2', '2026-10-09 14:30', 'none yet', 66.5)
('L4', '2026-09-20 10:00', '2026-09-21 10:00', 24.0)
('L3', '2026-10-01 11:00', '2026-10-02 09:00', 22.0)
('L5', '2026-10-11 16:00', 'none yet', 17.0)

Expected result: a short list of leads that waited too long, with the longest wait first. Common mistake: leads with no first touch yet. The query counts them as waiting until “now”, which is correct, so they surface as soon as they pass the threshold.

Step 3: Turn each gap into a trigger

A trigger has three parts: the event that fires it, the action it takes, and the condition that stops it. Writing all three prevents the two classic failures, silence and over-messaging.

Gap Trigger event Action Stop condition Owner
Confirmation depends on a person Form submitted Send a confirmation within minutes, every time None, one message Marketing
Callback depends on someone noticing No callback logged 24 hours after the form Alert the lead owner and log the alert Callback logged Sales manager
Follow-up depends on someone remembering Proposal sent Follow-up messages on day 3, day 10, and day 21 Prospect replies Sales rep
Hand-off depends on someone noticing the reply Prospect replies End the sequence and assign a rep Not applicable Sales rep
Four cards showing the triggers: form submitted sends a confirmation, no callback in 24 hours alerts the owner, proposal sent starts follow-ups on day 3, 10 and 21, and a reply ends the sequence
The four triggers from the table. The 24-hour and day 3, 10, 21 thresholds are examples to adjust to your sales cycle.

Most automation tools express this in two ways, and it helps to know which you are using. Event triggers fire when something happens, such as a form submission. Filter triggers fire when a record matches a condition. HubSpot’s documentation, for example, states that workflows can be triggered off filter-based criteria or event-based criteria, and that by default records enroll only the first time they meet the trigger unless you allow re-enrollment. That default matters for a proposal sequence, because a lead who reopens a deal would otherwise never re-enter it.

Screenshot of HubSpot Knowledge Base documentation on workflow enrollment triggers with two statements boxed and numbered: records are only enrolled the first time they meet the triggers by default, and workflows can be triggered by filter-based or event-based criteria
HubSpot Knowledge Base, “Set your workflow enrollment triggers” (page last updated 4 Sep 2026, read 11 Oct 2026). Box 1: default enrollment happens once. Box 2: event-based and filter-based triggers.

Step 4: Test the rules on sample events before connecting a tool

Logic errors are cheaper to find in a script than in a live automation. The script below implements the four triggers, reads a CSV of events, and prints the action due for each lead at a date you pass in. Giving it a fixed date makes the run repeatable.

journey_rules.py (run as python journey_rules.py events.csv "2026-10-12 09:00")

import csv, sys
from datetime import datetime, timedelta

FOLLOW_UP_DAYS = [3, 10, 21]                      # days after the proposal is sent
NOW = datetime.fromisoformat(sys.argv[2])         # pass "now" so the run is repeatable

events = {}
for r in csv.DictReader(open(sys.argv[1], newline="")):
    events.setdefault(r["lead"], []).append((datetime.fromisoformat(r["time"]), r["event"]))

def when(evts, kind):
    return next((t for t, k in sorted(evts) if k == kind), None)

for lead, evts in sorted(events.items()):
    kinds = {k for _, k in evts}
    actions = []
    form = when(evts, "form_submitted")
    if form and "confirmation_sent" not in kinds:
        actions.append("send confirmation now")
    if form and "callback_done" not in kinds and NOW - form > timedelta(hours=24):
        actions.append("ALERT: no callback within 24 hours")
    proposal = when(evts, "proposal_sent")
    if "replied" in kinds:
        actions.append("stop sequence, hand to a rep")
    elif proposal:
        for d in FOLLOW_UP_DAYS:
            if proposal + timedelta(days=d) <= NOW and f"followup_day{d}_sent" not in kinds:
                actions.append(f"send follow-up day {d}")
    print(f"{lead}: {'; '.join(actions) if actions else 'no action due'}")

The input is a log with one row per event. These are the illustrative events used for the test.

events.csv (illustrative)

lead,time,event
L1,2026-10-06 09:10,form_submitted
L1,2026-10-06 09:11,confirmation_sent
L1,2026-10-06 15:00,callback_done
L1,2026-10-07 10:00,proposal_sent
L2,2026-10-09 14:30,form_submitted
L3,2026-10-01 11:00,form_submitted
L3,2026-10-01 11:01,confirmation_sent
L3,2026-10-02 09:00,callback_done
L3,2026-10-02 16:00,proposal_sent
L3,2026-10-05 16:00,followup_day3_sent
L4,2026-09-20 10:00,form_submitted
L4,2026-09-20 10:02,confirmation_sent
L4,2026-09-21 10:00,callback_done
L4,2026-09-21 15:00,proposal_sent
L4,2026-09-24 15:00,followup_day3_sent
L4,2026-09-25 09:00,replied
L5,2026-10-11 16:00,form_submitted
L5,2026-10-11 16:01,confirmation_sent
L6,2026-10-12 08:30,form_submitted

Output on 2026-10-12 09:00:

L1: send follow-up day 3
L2: send confirmation now; ALERT: no callback within 24 hours
L3: no action due
L4: stop sequence, hand to a rep
L5: no action due
L6: send confirmation now
Timeline of the follow-up sequence after a proposal with markers at day 0, day 3, day 10 and day 21, and a note that a reply on day 7 cancels the day 10 and day 21 messages
The schedule behind the rules. A reply on day 7 means follow-ups 2 and 3 never send. Intervals are examples, not benchmarks.

Here is why each lead got the result it did, so you can check the logic against your own events.

Lead What the log shows Action due Why
L1 Proposal sent 7 Oct 10:00, no reply, no follow-up sent Send follow-up day 3 Day 3 was due on 10 Oct; day 10 is not due yet
L2 Form on 9 Oct, no confirmation, no callback Send confirmation; alert More than 24 hours with no callback
L3 Proposal 2 Oct, day 3 sent No action due Day 10 falls on 12 Oct at 16:00, after “now”
L4 Replied on 25 Sep Stop sequence, hand to a rep A reply ends the sequence
L5 Form 11 Oct 16:00, confirmation sent No action due Only 17 hours, under the 24-hour alert
L6 Form 12 Oct 08:30, no confirmation Send confirmation now The confirmation trigger has not fired yet

Common mistake: a rule that keeps firing after the prospect replies. The stop check runs first in the script for exactly that reason.

Step 5: Monitor and adjust

Once the rules run in your tool, review each trigger monthly. You are checking that it fires, that it works, and that it stops.

Check How to measure it What it tells you
Fired Count of times the trigger ran in the month Whether the trigger is wired up at all
Replied Prospects who responded within the sequence Whether the message and timing work
Stopped on reply Sequences that ended when a reply arrived Whether the stop rule works
Alerts actioned Alerts that led to a logged callback Whether the owner actually acts on them

Adjust one threshold at a time and compare the next month. Changing the callback alert and the follow-up days together hides which change mattered.

How do you fix the common problems?

Symptom Likely cause Fix
A prospect gets follow-ups after replying The stop condition is missing or checked after the send Check for a reply before every send, as the script does
A lead never enters the sequence Records enroll only once by default and this one already did Enable re-enrollment where a lead can legitimately restart
Alerts pile up unread No owner or too many alerts Assign a named owner and tighten the threshold
Timestamps are in different time zones CRM and email tool store different zones Convert everything to one zone before measuring gaps

Checklist

  • One real journey is written out stage by stage
  • Every gap is measured from your own data
  • Each gap has one trigger, one action, and one stop condition
  • The rules are tested on sample events before launch
  • Fired, replied, and stopped are reviewed monthly

Downloads: Journey trigger map template (CSV) and Journey automation checklist (PDF).

Where does this connect to AI agents and orchestration?

Journey automation is the strategy layer: which touchpoints exist, and what should happen at each one. The execution of those triggers is where AI orchestration and dedicated AI agents for sales come in. They are what actually sends the message, updates the record, or books the appointment when a trigger fires. Journey mapping without that execution layer is just a diagram; the automation is what makes the diagram run on its own.

The same logic extends past the sale. Once a lead becomes a customer, the trigger-and-action approach carries into onboarding and retention, covered in our guide to AI for customer success.

MV3 Marketing builds these journeys as part of its demand generation work, starting from a business’s actual customer path rather than a generic funnel template.

Layer What it does Where it lives in this guide
Strategy Decides which touchpoints exist and what each should trigger Steps 1 and 3: the journey map and the trigger table
Rules Encodes event, action, and stop condition Step 4: journey_rules.py, then your CRM workflow
Execution Sends the message, updates the record, books the meeting Your email tool, CRM, or an AI agent
Measurement Shows what fired, replied, and stopped Step 5: the monthly review

Frequently asked questions

What is customer journey automation?

Customer journey automation is the practice of connecting the individual touchpoints a prospect experiences, ads, website visits, emails, calls, so that each one can trigger an appropriate next action automatically, based on what the prospect did, rather than requiring a person to track and follow up on every lead manually.

How is this different from email marketing automation?

Email automation is one channel within the journey. Customer journey automation coordinates across channels, ads, website behavior, email, SMS, and calls, so the follow-up a prospect gets reflects everything they have done, not just whether they opened an email.

Where should a B2B SaaS team start mapping their journey?

Start with the actual path a real customer took recently, from first contact to signed contract, and write down every touchpoint and every gap where follow-up depended on someone remembering to do it. Those gaps are where automation has the most immediate impact.

Does journey automation require a large marketing team to manage?

No. Once the journey is mapped and the automated triggers are built, it runs continuously without daily manual oversight. The setup work is front-loaded; ongoing management is mostly monitoring and periodic adjustment.

If you have never mapped out where your leads actually go quiet, that is the fastest place to start. Book a call and we will map it with you.

Vance Moore
Vance Moore LinkedIn
Chief Growth Officer, MV3 Marketing

Vance Moore is Chief Growth Officer at MV3 Marketing and leads its team of SEO professionals, engineers and auditors.

Ready to audit your organic growth opportunity?

$2,500 flat. 5 business days. Six deliverables tied to pipeline , not rankings. No retainer required.

Get the Organic Growth Audit →

Turn Your Organic Channel into a Revenue Engine.

The MV3 SEO Audit maps your full organic opportunity in 5 business days.

Get the Audit →