A fresh lead is worth most in the hours after it arrives, and it loses value with every hour nobody calls. In our experience, many sales organisations know this and still lose leads, because their dialler gives up too early and their salespeople spend their day on people who were never going to buy. In one lead pre-selection campaign we ran, changing those two things raised the number of interested leads by 26% and cut customer acquisition cost by 73%.
This article describes how that campaign was designed and why the mechanisms behind it travel well. We do not name the client, and we are careful about what the numbers mean. They describe one campaign, at one point in time. The mechanisms are what you can use.
The situation#
The client worked in financial services, with a flow of leads from online enquiries: people who had expressed interest in a credit product and left a phone number. As in most such operations, leads arrived through the day in volumes that followed marketing activity, and a sales team was responsible for turning them into applications.
The pattern will be familiar to anyone who has run such a team. A form is filled in at lunchtime. The lead enters a queue. The dialler tries once or twice, reaches voicemail or a busy line, and moves on. By the time a salesperson actually speaks to the person, it may be the next day. By then they have compared other offers, received other calls, or simply lost the moment. And when the salesperson does get through, a large share of conversations are with people who are not interested any more, cannot be reached at a sensible time, or would not qualify for the product anyway.
Why the dialler gives up too early#
A conventional dialler in a call centre is tuned to keep agents busy. Its retry rules are a compromise between reaching people and not wasting agent time on unanswered calls, so it stops after a small number of attempts and spaces them in ways that suit the shift pattern rather than the customer. That is rational for a team paid by the hour. It is not rational for a lead that might answer on a later attempt, at a time that suits them.
Our position is simple: without reaching the customer, there is nothing else. Because the first contact is executed by Vee's AI rather than by agents, persistence does not cost anyone's working day. The calling strategy can be designed around the lead rather than around staff availability: with attempts designed around reaching the person, inside the windows and limits agreed with the client.
What we built#
The process had a narrow, clear job: find out quickly whether the person is still interested, collect the information the sales team needs, and pass the right people on at the right moment. Everything else was designed around that.
- Call while it is warm. New leads were picked up as they arrived and called promptly. Who calls first has an advantage, and in lead generation that advantage is measured in hours.
- Keep trying, intelligently. Unanswered leads were retried with more persistence than a dialler would allow, within the agreed limits, and the retry rules were tuned for this lead flow.
- Open in a way that keeps people on the line. The call introduced itself as AI, named the reason for calling (the person's own enquiry), and moved straight to the point. In cold and near-cold contact, studio-recorded voices often work better than synthesis, and voice was treated as a variable in the design rather than a default.
- Check current interest, not historical interest. The form said the person was interested yesterday. The call established whether they were interested now, and collected the basic information needed to judge whether a conversation with sales made sense.
- Status every record. Each lead ended with a clear status the client could act on: interested and transferred, interested with a callback time, not interested, unreachable after the agreed attempts, or an exception.
- Transfer live, within capacity. Interested people were connected to a salesperson during the sales team's working hours and only up to the number the team could handle. Outside those limits, a callback was scheduled instead.
Pacing transfers to the sales team#
The most common objection we hear in lead qualification is a fair one: will you flood our sales team? The answer has to be built into the process, not promised in a meeting. Live transfers were paced to the team's capacity and working hours. If every salesperson was busy, the interested lead was not left ringing in a queue; the callback was arranged for a specific time and the record updated.
This changes the shape of a salesperson's day. Instead of working through a list in which most calls go nowhere, they receive conversations with people who have just confirmed interest. Their time goes where their skill matters: explaining the product, handling questions, getting an application started.
Seeing the flow, not just the totals#
The client could follow the process at case level: how many leads had been touched, how many attempts each had received and with what result, which conversations ended in a transfer, and which records were still open. That visibility matters more in lead handling than in most processes, because the problems are about time. A spike in form submissions, a change in the share of unreachable numbers or a sales team running short of capacity shows up within hours, and the calling strategy can be adjusted the same day. Our team listened to selected calls daily, especially unusual ones, to hear what the statistics could not show: where people hesitated, which question confused them, which opening made them stay.
The result#
In this campaign, the share of leads that ended as interested contacts rose, and the cost of acquiring a customer fell sharply. Both figures come from one lead pre-selection campaign and are historical. They are not a rate to expect elsewhere; different lead sources, products and markets produce different numbers.
The two figures are linked. More interested leads came from reaching more people while they were still warm. Lower acquisition cost came from that, and from salespeople no longer spending their time on contacts who were not interested or not eligible. Neither effect required the salespeople to work harder; both came from changing what reached them.
What generalises, and what doesn't#
Three mechanisms carry over to almost any lead-qualification process.
- Speed to first contact. The value of a fresh lead decays over hours, not days. A process that calls within the same session as the enquiry starts from a different place than one that calls the next morning.
- Persistence without fatigue. More attempts, better spaced, reach people that a staffed dialler gives up on. This works only if the attempts stay within agreed limits and windows; persistence is not the same as pressure.
- Separation of qualification from selling. Salespeople should spend their time on people who are ready to talk. Qualification is a different job, and doing it at scale is where automation helps most.
One thing does not carry over: the quality of the base. We cannot jump over a bad base. If leads come from a source that produces few genuinely interested people, a faster and more persistent first call will find the interested ones sooner and waste less of the sales team's time, but it will not create interest that is not there. Different bases also need different measurement. A fresh lead from a form, a warmed contact such as a policy nearing renewal, a customer with a known profile and a worked-over recovery base each have their own realistic level of success, and a process should be judged against the right one.
Context matters too. In the third quarter of 2026, as banks in Poland tightened consumer-credit criteria, domestic credit-sales processes shrank across the market. A qualification process cannot offset a change in who is eligible to borrow. It can make sure that the sales team spends its time on the people who still are.
How we would measure it with you#
Numbers like these are only meaningful when both sides agree in advance how they will be counted. Before a lead-qualification process starts, we agree with the client what counts as an interested lead, which system is the source of truth for applications and signed contracts, the measurement window, and how results will be compared with the way leads were handled before. Interest is captured in our conversation; the contract is recorded in the client's systems, and that outcome has to come back to us to be matched to the case.
We also recommend starting on part of the lead flow rather than all of it. A controlled pilot on a sample, run in parallel with the existing approach, shows the effect on the same sources in the same weeks. It also lets the sales team adjust to receiving fewer, warmer contacts before the whole flow changes.
What to ask before you change your lead process#
- How long, on average, does a new lead wait for its first call today, and how many attempts does it get?
- What share of your salespeople's conversations are with people who are not interested or not eligible?
- How will live transfers be paced to your team's capacity and hours, and what happens when everyone is busy?
- Which statuses will come back for each lead, and will they update your CRM?
- How will the result be measured: interested leads, applications, signed contracts, or cost per acquired customer, and against what?
If the honest answer to the first question is "the next day", the lead has already cooled. That is usually the first thing we change when we run a qualification process for a client.
Follow The First Second
Occasional issues on the mechanics and economics of AI-powered customer contact. No noise, no vendor news.
Follow via RSS