Fischer Future Heat texted week-old leads with AI and booked 105 home surveys

Setterly's AI sales agent booked 105 qualified home surveys in its first 30 days for UK electric-heating company Fischer Future Heat, which routes every enquiry still unbooked after seven days to the agent, according to a Setterly case study. The agent texts as "Claire from Fischer," asks three qualifying questions and books eligible homeowners into the survey calendar, producing 19 surveys on day one and 50 in week one; Setterly says the system went live on 15 September 2026 and Fischer paid per booked survey. No control group, show rate or close rate has been published, and the £630,000 pipeline figure is 105 appointments multiplied by an average order value of roughly £6,000.

10 min read Fischer Future Heat let an AI sales agent from Setterly text every enquiry still unbooked after a week. Setterly reports 105 qualified home surveys in the first 30 days. TL;DR - Fischer Future Heat, a UK electric-heating company, hands every enquiry still unbooked after seven days to Setterly’s AI sales agent, which texts as “Claire from Fischer”, asks three qualifying questions and books eligible homeowners into the survey calendar. - Setterly reports 19 surveys on day one, 50 in week one and 105 qualified home surveys in the first 30 days. - No control group, show rate or close rate has been published, and the £630,000 pipeline figure is 105 appointments multiplied by an average order value of roughly £6,000. - The copyable lever is the boundary. A timer picks the leads, fixed questions qualify them and a calendar completes the action, so the AI works inside a small, safe workflow. Every sales team has a pile of enquiries that nobody is chasing any more. The people asked, the team followed up, nothing happened, and newer leads took over the queue. Fischer Future Heat decided that pile deserved a second look. Four hours after the new texting system went live, its head of marketing asked whether the appointments were real. By the end of that first day, 19 home surveys had landed in the calendar. The interesting part is where Fischer put the AI. The UK electric-heating company let its own team work every enquiry first, and only a lead still unbooked on day seven entered the Fischer Future Heat AI lead recovery test. That boundary turned a vague automation project into a clear growth question, which was whether a short automated conversation could win back demand after the normal sales process had run its course, without crowding fresh leads. About 2,000 warm leads a week went quiet On its company website https://fischerfutureheat.com/ , Fischer Future Heat sells electric heating systems to UK homeowners. Its funnel, the path a customer follows from first contact to purchase, begins with an enquiry and moves towards a home survey, where the company assesses the property and discusses a system. According to the case study https://www.setterly.co/results/fischer-future-heat from Setterly, the company behind the AI agent, Fischer generates around 2,000 enquiries a week that are still unbooked after seven days. These were warm leads. Each one was a person who had asked about improving their heating and then stopped moving. Marketing had paid to create that demand and sales had tried to turn it into an appointment, so the gap sat between the two. Any growth team will recognise the pattern. Leads stop getting attention once newer demand arrives, and the usual answer is to buy more acquisition. Fischer tested whether the cheaper lever was to change the follow-up system around demand it had paid to create. Three questions on day seven The intervention was narrow. On day seven, every enquiry still unbooked moved to Setterly, whose AI sales agent sent an SMS as “Claire from Fischer” and asked three things, whether it had reached the right person, whether that person owned the home and whether they were still considering a heating upgrade. If the answers fitted Fischer’s rules, the agent booked a survey straight into a calendar managed by the appointments team. Fischer paid per booked survey, and Setterly says the system went live on 15 September 2026. The structure worth borrowing is the seven-day lead handoff. Fischer’s sales team kept the first seven days, and Fischer’s own rules decided what counted as a qualified customer. A timer picked the eligible leads, fixed questions collected the required details, business rules decided qualification and a calendar completed the next action, so the language model worked inside a bounded workflow. That makes AI SMS lead follow-up copyable without a machine learning team. A company needs a trigger, a short approved set of questions, explicit qualification rules, a link to its scheduling tool and a handoff to a person for anything outside the script. The hard design choice is where the boundary sits, and the choice of model matters far less. Reading the first month of AI lead recovery Setterly reports 19 surveys on the first day and 50 in the first week, reaching 105 qualified home surveys after 30 days. Fischer’s average order is roughly £6,000, and Setterly multiplied the 105 bookings by that figure to describe £630,000 in pipeline, meaning potential sales still to be won. The appointment count is the cleaner number to track. It is an observable step in the funnel and one the growth and sales teams both own. The £630,000 is a planning estimate that sits some way from booked revenue, because it assumes every survey carries the full average order value and the public case does not yet say how many appointments were attended, quoted or closed. A few more figures would turn the 105 into a conversion rate. Setterly says around 2,000 day-seven leads enter the system each week, and the case does not give the exact number contacted in the 30-day window, the reply rate, the opt-outs or how many people failed qualification. Marketers call that missing total the denominator, the number the 105 is measured against, and it is what turns a count into a rate. Counting the extra bookings The case supports a clear claim. Setterly’s system booked 105 surveys from leads that had not booked during Fischer’s first seven days of follow-up. What remains open is how many of the 105 were extra, a question growth teams call incrementality, meaning the lift you would not have got otherwise. There was no randomised holdout, which is a group of leads kept back from the new treatment so you have something to compare against. Some homeowners may have returned on their own, answered a later call from the team or booked through another channel, and weather, pricing, campaign mix and the age of each lead could all have changed response. Setterly built the system and published the results, and no independent audit has been disclosed. Cann’s holdout https://industrycontents.com/applovin-incrementality-test-cann-geo-holdout/ shows what that comparison looks like in practice, and Matas https://industrycontents.com/ai-retention-experiment-matas/ used a control group to test whether targeted outreach kept quiet subscribers. That missing comparison matters more here than in typical AI appointment setting, because a booked appointment can look valuable while shifting work elsewhere in the funnel. If low-intent bookings turn into no-shows or weak close rates, the calendar fills and revenue stays flat. Following each lead through attendance, quote and paid order would settle it. Two recent Growth Signal cases turn on the same distinction. Atlas24h https://industrycontents.com/atlas24h-ai-sales-agent-revenue/ connected a conversational agent to confirmed contracts, but had no holdout for incrementality. Napster’s concierge https://industrycontents.com/belgian-grand-prix-ai-concierge/ for the Spa Grand Prix moved closer to checkout, while its reported revenue remained an estimate. Fischer sits earlier in the journey, since its public evidence stops at the appointment. A cleaner test for stale lead reactivation A clean follow-up test would randomise eligible day-seven leads before any automated message goes out, with half receiving Setterly’s conversation and half continuing under the existing process. Closed revenue per eligible lead after a fixed window would be the primary metric, and completed surveys per lead would be the early signal along the way. Diagnostic metrics would cover delivery, reply, qualification, booking, attendance, quote and close rates, with opt-outs, complaints and human intervention watched as guardrails, the measures that warn you when a gain is costing something. Splitting results by lead source and lead age would show whether stale lead reactivation works because of the channel, the timing or the quality of the original demand. If the lift holds up, Fischer can then vary the handoff point. Trying day three, day seven and day fourteen would show when the expected recovery value starts to outweigh the risk of overlapping with the human team, which is a more useful question than whether another prompt produces a friendlier message. Fischer’s first month suggests that the forgotten edge of the funnel can still produce appointments. The next experiment can show how many of those appointments would otherwise have disappeared, and how many end up as paying customers. | Experiment detail | What the reporting shows | |---|---| | Constraint | Around 2,000 enquiries a week remained unbooked after Fischer’s sales team had worked them for seven days. | | Intervention | Setterly’s AI SMS agent contacted day-seven leads, asked three qualifying questions and booked eligible homeowners directly into Fischer’s survey calendar. | | Reported result | Setterly reports 19 surveys on day one, 50 in week one and 105 qualified home surveys in the first 30 days. | | Evidence | Named UK customer, named marketing practitioner and dated production launch. Setterly built the system and published the case. | | Limits | No control group, contact count, response rate, show rate, close rate or audited revenue has been published. The £630,000 pipeline figure is 105 appointments multiplied by an approximate £6,000 average order value. | Sources and disclosure - Setterly case study https://www.setterly.co/results/fischer-future-heat of Fischer Future Heat, covering the first 30 days after the 15 September 2026 launch - Fischer Future Heat company website https://fischerfutureheat.com/ and its profile https://fischerfutureheat.com/blog/author/tnurserangemarketing-co-uk/ of Tom Nurse - Tom Nurse https://uk.linkedin.com/in/tom-nurse-93775079 and Setterly founder Cameron Richards https://uk.linkedin.com/in/camarrich on LinkedIn - Fischer Future Heat UK Limited at Companies House https://find-and-update.company-information.service.gov.uk/company/06021954 Setterly built the system and published the case, so the survey counts and the pipeline estimate come from the supplier. They have not been independently audited, and the public account gives no control group, contact count or close rate. FAQ What did Fischer Future Heat change? It sent every enquiry still unbooked after seven days to an AI texting agent, which asked three qualification questions and booked eligible homeowners into a survey calendar. Did the experiment generate £630,000? Setterly describes £630,000 as pipeline, worked out by multiplying 105 booked surveys by Fischer’s average order value of roughly £6,000. Attendance, close rate and paid revenue have not been published yet. Was there a control group? No. Setterly has not published a randomised holdout, so it is unclear how many of the 105 surveys would have been booked without the AI texts. How many leads reach the AI agent? Setterly says around 2,000 enquiries a week are still unbooked after seven days. The case does not give the exact number contacted during the first 30 days. What can another growth team copy? Pick a clear point where normal follow-up ends, send only that group of leads into a short automated conversation, connect the qualification to a real next step such as a calendar booking, and measure closed revenue against a random holdout. Compare more measured experiments in the Signal Index https://industrycontents.com/ic-lab/growth-signal-index/ , browse the Growth Signal https://industrycontents.com/category/growth-signal/ archive, or use the benchmark tools in the IC Lab https://industrycontents.com/ic-lab/ .