About 860 new enquiries a week.
Every one needed a reply, a few qualifying questions and, if it fitted, a call booked with the right person.
Case study · Built and run by our team
An education business.
$5.8M+ won from AI-booked calls.
An education business serving the UK and US had more enquiries than its team could follow up personally. We built, and still run, the AI SMS system that now has the first conversation, qualifies interest and books calls with its team.
Read this first.
This is our own previous work, not a Second Look client, and not UK results. The results below come from the business’s US programme, April to October 2026. These were mostly new enquiries followed up quickly, not aged leads. They show the engine works; they are not a forecast for your leads.
The challenge
More enquiries than
people to answer them.
53% of threads ended unanswered.
In a review of 369 SMS conversations before launch, more than half ended with the lead’s last message never answered.
Every rep did it differently.
Before any automation, the team replied by hand. We ranked those replies by what actually booked and built the best of them into the system.
What we built
The first conversation,
handled properly.
Qualify, then book
The AI answers questions, checks fit and offers real slots from the team’s calendars, only at sensible hours in the lead’s own time zone.
People where it matters
Complaints, sensitive questions and anything outside the brief pause the AI and go to a named person on the team.
Opt-out comes first
Every inbound message is checked for opt-out before anything else, and anyone who asks to stop is suppressed straight away.
The results
Six months,
in numbers.
US programme, 18 April to 17 October 2026. Counted from the system’s own message and booking records.
22,000+leads contacted22,352 messaged · 29% replied
$5.8M+won from calls our AI booked183 won deals. $840k+ where the AI booked the first call
2,000+calls booked by the AINo human needed to book. People handle escalations, complaints and supervision.
61%show rate on AI-booked callsvs 45% for calls booked other ways
Revenue is the deal value at win (list price), not cash collected. Of the $5.8M+, $843,799 came from 26 deals where the AI booked the first call; $5.0M came from 157 deals where the AI scheduled a later-stage call for a lead already in the pipeline, and the team closed. Recent bookings have not had time to close, so these are floors. About 3% of contacted leads opted out. AI-booked leads had already replied, so the show-rate gap is not a controlled comparison.
What we learned
Three lessons we
bring to your pilot.
The conversation is what lifts attendance.
Calls booked through the SMS conversation showed up about 61% of the time. Calls booked through a self-serve web booking page and calls booked by hand were both about 47%. Automation alone did not help; the conversation did. (April to June 2026 analysis.)
The first reply decides most bookings.
Leads who replied to the opening messages produced almost all bookings. Leads who first replied later produced very few, so follow-up has to be quick and well judged, not endless. (June 2026 analysis.)
Guardrails are the product.
Daytime-only slots, opt-out first, human escalation and one live offer at a time. Each rule came from a real conversation that went wrong, and each one is in the Second Look pilot from day one.
What it means for you
The same engine,
pointed at your aged leads.
This case study proves the conversation, the booking and the show rate. A founding pilot tests what it hasn’t yet proved: reactivating aged UK enquiries in your sector.
Pay for qualified bookings the lead confirmed. No-shows replaced free.
Check your fit ↗