Example Scenarios

The situation

A three-provider med spa runs Instagram ads and posts before-and-after results. The enquiries arrive as DMs, mostly between 8pm and midnight, and mostly asking the same two questions: how much is it, and when can I come in.

The front desk answers them the next morning between clients. By then a good share of those people have messaged two other spas as well. Nobody can say how many were lost, because a DM that never gets a reply leaves no trace in any report.

What gets built

• AI chat on Instagram, Facebook and the website — One assistant answering all three, in English or Spanish, using price ranges and treatment information the spa has approved.
• Consultation booking inside the conversation — Live provider availability with correct treatment durations, so the enquiry ends in a confirmed appointment rather than a promise to call back.
• Confirmation and preparation messages — Sent automatically, including what to avoid before the appointment and where to park.
• Reminder and no-show recovery — A reminder timed to the spa's own no-show pattern, and an automatic rebooking offer within minutes if someone doesn't arrive.
• Escalation rules — Anything clinical — suitability, medications, contraindications — is routed to a provider rather than answered by the assistant.

What changes day to day

• Evening enquiries get a reply in seconds instead of the next morning.
• The front desk stops re-typing the same two answers all day.
• Consultations are booked during the conversation, not after a phone tag.
• Missed appointments generate a rebooking attempt without anyone remembering to make one.
• Every enquiry is recorded against a contact, so the spa can finally see how many there were.

How we'd measure it

Before anything is switched on we record the baseline, so the difference is measured rather than asserted. For this business that means:

• Enquiries received per channel, per day and per hour
• Time from first message to first reply
• Enquiry-to-consultation booking rate
• No-show rate before and after reminders
• Consultations that came from outside opening hours

The situation

A six-technician air conditioning company. The owner runs the phone between jobs, and in summer it rings constantly. Calls come in while crews are working, while the owner is driving, and at night when a system fails in the heat.

Voicemail is effectively a dead end — most callers hang up rather than leave one, and by the time anyone listens, the customer has called the next company on the search results page.

What gets built

• Missed call text back — Any unanswered call gets a text from the business number within seconds. Most people reply, because replying to a text is easier than calling again.
• After-hours AI receptionist — Night and weekend calls are answered, the problem and address are captured, and the job is either booked or escalated to whoever is on call.
• Emergency rules the owner defines — No cooling, water leak, no power — these route to a person immediately under rules written down in advance rather than judged case by case.
• Bilingual handling — Calls and texts handled in English or Spanish, detected automatically.
• Quote follow-up and review requests — Estimates chased on a schedule, and a review request sent the moment a job is marked complete.

What changes day to day

• Calls that used to go to voicemail become text conversations the same minute.
• The owner stops being the answering service between jobs.
• Night and weekend emergencies reach an on-call technician instead of a voicemail box.
• Quotes get chased consistently rather than when someone remembers.
• For the first time the business can see how many calls it was actually missing.

How we'd measure it

Before anything is switched on we record the baseline, so the difference is measured rather than asserted. For this business that means:

• Missed call volume per week — the number most owners have never seen
• Reply rate to the missed-call text
• Jobs booked from calls that were originally unanswered
• After-hours calls captured versus lost
• Review count before and after automation

The situation

A four-agent team buying leads from portals and running Meta ads. Portal leads are sent to several agents at the same moment, so the conversation goes to whoever gets there first.

The agents are showing property for most of the day, which is exactly when the leads arrive. Long-cycle buyers — the ones looking to move in six or nine months — get two calls and are then forgotten entirely.

What gets built

• Instant response to every lead source — Portal, website, Meta and Google lead forms all feed one system that sends a real, personalised message within seconds.
• AI callback within the first minute — For higher-value enquiries, a voice agent calls while the person is still on the listing.
• Qualification the team defined — Budget, timeline, financing status, areas and property type — asked consistently, recorded on the contact, and reviewed by the broker to stay well clear of fair-housing risk.
• Agent routing — By area, price band, language or availability, so a Spanish-language enquiry reaches an agent who can handle it immediately.
• Long-cycle nurture — Buyers who are months away stay in contact automatically until they're ready, instead of going cold after the second call.

What changes day to day

• Every portal lead gets a reply in seconds, including during showings.
• Agents spend their first conversation on qualification that's already been done.
• Enquiries in Spanish reach the right agent without a delay.
• Nine-month buyers stay in the pipeline instead of quietly disappearing.
• Campaign data survives all the way to the booked viewing, so the team can see which lead source actually produces business.

How we'd measure it

Before anything is switched on we record the baseline, so the difference is measured rather than asserted. For this business that means:

• Median time from lead received to first contact attempt
• Contact rate per lead source
• Leads that reached a booked viewing
• Cost per booked viewing by campaign
• Long-cycle leads still engaged after 90 days