Most small business owners spend hours every week on tasks that could run themselves: following up with leads who went quiet, checking in with clients after an appointment, reminding happy customers to leave a review. It's not complicated work but it eats time that should go toward actually growing the business.
We built a system that handles all of it. The result: the business owner now spends 3–10 hours a month on operations that previously demanded full-time attention, and Google reviews increased by 70%.
Here's exactly how it works.
The Problem We Were Solving
The business had three recurring manual tasks:
- Post-appointment follow-up: existing clients who'd just completed a session or purchase. No structured touchpoint meant no repeat business signals.
- Google review collection: the team knew happy clients would leave reviews if asked, but asking manually felt awkward and inconsistent.
All three were being handled reactively and inconsistently or not at all.
The System We Built
Stack
- Python: orchestration, scheduling, API calls
- Claude AI: message generation, tone calibration
- Existing CRM/booking data: trigger source
Cold Lead Sequence (4 Touchpoints)
When a lead enters the pipeline without booking, the system initiates a 4-message sequence spaced over days, not hours. Each message is generated by Claude AI using context about the lead (service they enquired about, how they came in) so it reads like a personal message but not a template blast.
The tone shifts across the sequence: curious → helpful → specific → final. No pressure, no repetition. The fourth message is a clean close that leaves the door open without pestering.
Post-Appointment Follow-Up (4 Touchpoints)
When an existing customer completes an appointment or purchase, a separate 4-message sequence fires. These messages do different jobs:
- Confirmation + gratitude: sent within hours of the appointment
- Check-in: 2-3 days later, asking how things went
- Value add: a tip, resource, or insight relevant to their service
- Review request: timed for when satisfaction is highest
The review request is the fourth message, not the first by that point, goodwill is built and the ask feels natural.
Google Review Collection
The review request message links directly to the business's Google review page one tap, no friction. No asking the client to "find us on Google." A direct link to the exact review form.
This single change, combined with the timing (sent at peak satisfaction), drove a 70% increase in Google reviews.
What the Owner Actually Does
Once the system is live, the owner's involvement is:
- Review new leads as they enter (they see a summary, not raw messages)
- Occasionally adjust a message sequence if a service changes
- Check a weekly report
That's 3–10 hours a month. Everything else runs without them.
Why This Works
The system works because it does things humans consistently skip under pressure:
- It follows up every time, not just when the team remembers
- It waits the right amount of time between messages
- It sounds human because it's written by Claude AI with actual context, not a mail merge
- It asks for reviews at exactly the right moment
Most businesses lose reviews and repeat business not because clients are unhappy but because no one asked, or asked at the wrong time.
What You Can Automate in Your Business
If your business has any of these, you're a candidate for the same system:
- A booking or appointment flow
- Leads that come in and go quiet
- Happy clients who never left a review
- A CRM you update manually
The technical stack are Python, Claude AI. /Your existing tools can be adapted to most service businesses without rebuilding everything from scratch.
Want to see what this would look like for your business? Send us a message on Instagram we'll map out which workflows make the most sense to automate first.