B2B database reactivation case study: AI won back $200K+ in lost deals
A B2B fleet telematics company used AI database reactivation on its closed-lost deals and won back more than $200K, about 35 times the cost.
An AI sales agent at the top of the funnel answered every new lead for a rental manager, cleared out the dead ones and signed $495K, a 17x return.
A national short-term rental manager was losing property owners to whichever competitor rang them first. Since an AI agent took over first contact, 82% of new leads reply to it, leads it books are reached by the sales team far more often and sign at more than twice the rate, and $495,000 in lifetime revenue has signed through the calls it booked. That is about 17 times what the agent costs.
An owner had been running her own holiday rental for two years. She was tired of doing it herself and tired of cleaners who did not turn up, so late one Friday night she filled in an enquiry form with a property manager. Owners like her usually ask two or three managers at once.
Four minutes later her phone buzzed. Names and the address are hidden.
Agent: Hi [name], it is Zara from [business]. Just saw your enquiry come through and wanted to confirm a couple of details before I get you across to one of our specialists. Is that alright?
Owner: Hi Zara ofcourse
Agent: Just confirming that’s [address] you’re looking to put on short-term rental?
Owner: Yes correct, it’s already on air BnB and I have been running it myself for almost 2 years however it’s consuming a lot of my time and my cleaners aren’t very reliable so I would like a property manager to just manage it for me :)
Agent: Unreliable cleaners on top of running it yourself is a lot. How long would you be looking to have it managed for?
Twenty-eight minutes after she enquired, she had a call booked with a specialist for Monday afternoon. None of it needed anyone at the company.
The company manages holiday rentals for property owners around Australia, and its marketing lead knew exactly how the business is won: “the first person to pick up the phone and call generally wins the contract.”
The trouble was when owners enquire. Nearly half of all enquiries arrived outside office hours, and more than half of the texts owners sent came in the evening, overnight or on a weekend. During the day the sales team was busy running calls and onboarding. Covering every enquiry at the speed the market demanded meant, in the marketing lead’s words, “a whole team of people just booking meetings.”
So leads waited. Evening and weekend enquiries heard back the next working day, and by the company’s own measure it took three to four days to get a lead into a real conversation with a salesperson. By then, some of them had already gone with someone faster.
We built an agent that texts every new lead shortly after they enquire, at any hour, under a persona name the team chose. It confirms the property, asks how they want it managed, answers questions about fees, cleaning and furnishing, and books a call straight into a specialist’s calendar.
Two design choices mattered as much as the speed. When a lead books, the agent emails the rep the whole conversation and steps back, so the lead only ever hears from one person. And when a lead is not a fit, such as someone who only wants cleaning or can only rent for three months, it says so politely and closes the conversation before a rep spends a minute on it.
The agent is not a separate tool the team has to check. It is connected to the company’s CRM and works as part of the sales process, and the order of play is simple: the agent gets to every new lead first, and the team takes over the moment it is useful.
The result is a sales team that spends its day on booked calls and warm leads, while the agent handles first contact, qualifying, chasing and clearing out.
82% of leads replied to the agent, and almost half of those replied within five minutes of the first text. Across the same period, the sales team got to speak to 60% of all leads. Most businesses we talk to would be thrilled with either number, and in our experience the difference is speed: a text that lands while the lead is still holding their phone tends to get answered. Research backs that up. A Harvard Business Review study found that firms that tried to contact a lead within an hour were nearly seven times as likely to qualify it as firms that waited even an hour longer.
The agent went live in late March. In its first full month it booked a call for almost half of everyone who enquired, and from then on the booking rate held between 38% and 49% every month, through a quiet winter and a busy spring.
Yes. Leads the agent booked signed at more than twice the rate of the rest, and 92% of the deals that signed had talked to the agent first. Booked calls are only worth something if they turn into signed deals, and in November the team wanted to know whether the agent’s bookings did. So we read every enquiry since launch against the company’s own CRM, deal by deal.
The answer was clear, and three numbers explain it.
The sales team reached more of the leads the agent booked. Reps got through to 79% of the leads the agent had booked a call for, against 46% of the leads it had not. A booked call is a call the lead is expecting, so far more of them actually happen.
Booked leads signed more than twice as often. Of every lead the agent booked, 8.5% became a signed deal. Of every lead it did not book, 3.7% did. Part of that gap is the call-through rate above, so we also compared only the leads a rep actually spoke to. Booked leads still signed more often, 10.7% against 8.1%. The booking got more leads onto a call, and the leads on those calls were more likely to sign.
The agent was there at the start of almost every sale. 92% of the deals that signed came from leads who had talked to the agent first.
At the company’s own lifetime value per property, the deals that signed after the agent booked the call come to $495,000.
The sales lead summed up the speed difference: leads were moving from enquiry to a booked call “almost 10 times faster”, and “keeping [the agent] going is like that’s a dead set. I don’t think anyone’s going to argue that we should drop it.”
Yes, and this campaign shows why. About 8% of the deals that signed came from leads who never replied to the agent at all. The reps got them on the phone and won them.
The sales team never sat back and waited for bookings either. When a lead had not booked, a rep would often ring them anyway, and if the lead did not pick up, the agent kept the conversation going by text until they replied. Many of the deals that signed without an agent booking still started with the lead replying to the agent before a rep called.
So the booking is one goal. The bigger one is reaching every lead as fast as possible, by whichever route gets an answer first. The agent is an extra force on top of the team’s calls: it answers at any hour, never waits until morning and keeps following up, and the reps spend their time on the conversations it has already warmed up.
We run the agent rather than hand it over, so the work did not stop at launch. In November we cut the wait before the first text to about three minutes, and every rep now gets a weekly list of who the agent booked for them, who still needs a call and who needs a follow-up.
The data also gave the team something specific to look at: weekend enquiries book a call more often than any other group but sign less often once they reach a rep.
| Result | Figure |
|---|---|
| Lifetime revenue signed through calls the agent booked | $495,000 |
| Return on the agent | About 17 times its cost |
| Leads who replied to the agent | 82%, almost half within five minutes |
| Leads the sales team spoke to, across all enquiries | 60% |
| Call-through rate on leads the agent booked, against the rest | 79% against 46% |
| Signed deals, leads the agent booked against the rest | 8.5% against 3.7% (10.7% against 8.1% counting only leads a rep spoke to) |
| Signed deals where the lead talked to the agent first | 92% |
| Time to a real sales conversation | Under two days for leads the agent handled, against three to four days for the rest (the client’s own measure) |
| Enquiries outside office hours | 48% |
| Enquiries cleared from the call list (not a fit, not interested, or asked to stop) | About 1 in 5 |
| Leads ruled out as not a fit that the team later agreed were not a fit | All of them |
| Team time saved | About 150 hours of first-contact work |
How these are worked out: enquiry, booking and signing figures come from the agent’s conversations and the client’s CRM, 27 March to 13 November 2026. Revenue uses the client’s own lifetime value per property. Time saved assumes about ten minutes of a person’s time per enquiry for the first contact, the qualifying questions and the booking. The agent also worked the company’s old enquiries, and the client told us that brought in extra deals from leads it would never have gone back to.
It is if your enquiries arrive through a form and your team rings them back when it can, because you are in the same race this company was losing. Keep calling your leads. Then add something that answers every one of them within minutes, at any hour, and keeps following up until they reply.
You can see how the agent works on our speed to lead page and our out-of-hours page, and read more threads like the one above on our real conversations page. Most Australian businesses we tested took hours to reply. To find out where your own enquiries go quiet, book a free revenue engine audit.
For this short-term rental manager, the deals that signed after the agent booked the call came to $495,000 in lifetime revenue, at the company's own lifetime value per property. That is about 17 times what the agent cost.
In this campaign they did. Of every lead the agent booked, 8.5% became a signed deal, against 3.7% of the leads it did not book. Comparing only leads a rep actually spoke to, booked leads still signed more often: 10.7% against 8.1%.
Yes. Here about 8% of the deals that signed came from leads who never replied to the agent, and the reps won them by phone. The agent adds speed and persistence on top of the team's calls, it does not replace them.
We map where your leads go quiet, then show you the conversations an agent would have had. Thirty minutes, no obligation.