Speed to lead: call every new lead within minutes

Every sales team agrees that fast follow-up matters. Almost none of them do it. That gap is the whole opportunity, and it exists for a structural reason rather than a motivational one: the person who should make the call is busy doing something else when the form comes in.

What the speed-to-lead research actually says

The statistics you see quoted everywhere trace back to a small number of studies, and several are repeated with numbers nobody can source. Here is what holds up.

The five-minute rule. The Lead Response Management study, run by Professor James Oldroyd with InsideSales.com data, covered three years of web leads from six companies: more than 15,000 leads and over 100,000 call attempts. Its headline finding was a 21-fold drop in the odds of qualifying a prospect when response time stretched from 5 to 30 minutes, and a fourfold drop between 5 and 10 minutes. The likelihood of making contact fell more than tenfold within the first hour.

Two honest caveats. The study measured contact and qualification, not revenue. And it is old. Buyers in 2026 screen unknown numbers harder than they did then, which, if anything, makes the window more important: a lead who just submitted your form is expecting a call, and that expectation fades.

Almost nobody is fast. When Workato submitted demo requests to 114 B2B companies (the write-up on its site is dated March 2026), more than 99% did not respond within five minutes. Not one company called inside that window. The average time to a personalised email was 11 hours 54 minutes, the average time to a phone call was 14 hours 29 minutes, and only 31% of the companies called at all.

Read those two studies together and the conclusion is uncomfortable but useful. The benchmark is five minutes. The market average is measured in hours. A business that reliably calls within five minutes is not slightly better than its competitors; it is in a different category.

Why teams miss the window

Nobody decides to call leads tomorrow. It happens because of how the day is built.

• The lead arrives while the rep is on another call. The form notification sits behind the conversation they are already having.
• Leads arrive in clusters. Ad campaigns and email sends produce bursts. Ten leads in twenty minutes means the tenth waits for the first nine.
• Evenings and weekends. A homeowner fills in your quote form at 9pm. Nobody is there until morning, and by then they have filled in two more.
• Routing takes time. Round-robin rules, territory checks and CRM enrichment add minutes before a human even sees the lead.
• The first call is a chore. It is mostly voicemail, repeating the same four qualification questions, and booking a slot. It is the least rewarding call of the day, so it slips.

Every one of those is a capacity problem. Hiring a faster rep does not fix a burst of ten leads, and it does not fix 9pm.

The response pattern that wins

The goal is not to have an AI close your deals. It is to make sure every lead gets a fast, competent first conversation, and that your team spends its time on the people who are ready.

1. 1

   The form fires a webhook

   Your form backend, CRM or automation tool posts the new lead (phone, name and anything useful, such as the service they asked about) to a Dial Milo campaign's secret "start a call" URL. Include the form's submission ID and a re-sent submission is ignored rather than called twice.

2. 2

   Milo calls within minutes

   The call goes out as soon as a line is free, provided it is inside calling hours in the lead's local time. Milo says it is an AI, mentions the enquiry they just made, and checks it is a good moment.

3. 3

   It qualifies with your questions

   Three or four questions you would ask anyway: what they need, by when, where, rough budget. Answers about your services come from your script and the knowledge you upload, never made up.

4. 4

   It books or schedules the next step

   With Google Calendar connected, Milo can offer real free slots and book one. If the lead would rather speak later, it records a callback request with the time.

5. 5

   Your team picks up a warm lead

   The transcript, summary and outcome are recorded on the call, so whoever follows up starts from what was already said rather than from "so, what can we do for you?"

If the lead misses the call and rings the number back, Milo recognises them and carries on the same conversation. That matters more than it sounds: returned calls are some of the warmest you will ever get.

What the first call should sound like

The first call has one job: turn a form submission into a booked next step, quickly and politely. Here is an example for a fictional roofing company:

Milo: Hi, is that Dan? This is Milo, an AI assistant from Northside Roofing. You asked for a quote on our website a few minutes ago. Is now a good moment for two quick questions?

Dan: Yeah, go on.

Milo: Is this a repair or a full replacement, and roughly how soon do you need it?

Dan: Repair. There's a leak over the kitchen, so soon.

Milo: Understood. Our surveyor can come out on Wednesday at 9:00 or Thursday at 2:00. Would either work?

Notice what it does not do. It does not pitch, list services or ask for a budget before it has earned the right to. It references the enquiry so the call is clearly expected, asks the two questions that decide the next step, and offers real times.

After-hours leads and calling hours

Fast does not mean any time. In the US, the FTC's Telemarketing Sales Rule restricts telemarketing calls to between 8am and 9pm at the called person's location unless they have consented otherwise, and good manners usually argue for a narrower window than that.

So a lead that arrives at 11pm should not get a phone call at 11:01pm. With Dial Milo, the call waits in the queue and goes out at the start of calling hours in the lead's own time zone, which you set per campaign (the default is 10:00–20:00). If your leads tend to arrive late in the evening, consider starting the calling window earlier, within the law, so they are first in the queue the next morning, ahead of every competitor who checks email at 9:30.

Consent comes first

If you are calling web leads in the US with an AI voice, read this twice. On 8 February 2024 the FCC ruled that AI-generated voices are "artificial" under the Telephone Consumer Protection Act. The FCC's consumer guide spells out that prerecorded telemarketing calls need prior written consent, which can be given through a website form, and that AI-generated voice calls are illegal unless the consumer has agreed or the caller is exempt.

In practice: put clear consent wording beside the phone field on your form, saying the person agrees to be called about their enquiry, including by an AI voice. Store the record. And honour "stop calling" immediately; Dial Milo records it as an opt-out and keeps the number off future campaigns. The wider picture is in our guide to AI outbound calling.

What it costs to call every lead fast

Worked example (illustrative figures). Say a home-services company gets 200 web leads a month from US numbers, and today the average first call happens about three hours later.

• Assume 60% answer the AI call (people who just filled in your form answer more often than a cold list), so 120 conversations.
• Assume 3 minutes each. The line bills 3 started minutes at 4¢: 120 × 3 × 4¢ = $14.40. The AI bills talk time at 9.57¢: 120 × 3 × 9.57¢ = $34.45.
• The other 80 leads are mostly voicemail or no answer and are retried. Budget about 5¢ per voicemail attempt: say 120 attempts, $6.

Total: about $55 a month to give every lead a first conversation within minutes. Whether that is worth it depends on one number only you know: how many extra jobs a month you need to win to cover it. For most businesses the answer is a fraction of one.

You can price any number on the rates page and plan a list with the spend calculator. There is no subscription; calls that never connect have no minute charge.

How to measure your speed to lead

Before changing anything, measure where you are. It takes an afternoon.

Use the median, not the average. Two leads that waited a weekend will drag an average up and hide how fast you are on a normal Tuesday, or how slow.

Then run the AI on half your leads for a month and compare lead-to-booking against the other half. That is the only honest test of whether speed to lead moves your numbers, and it costs very little to find out.

The point most speed-to-lead advice misses

Speed only helps if the first conversation is good. A fast call that cannot answer "do you cover my area?" or "how soon can you come?" just burns the lead sooner. Give the AI the four things a good first call needs: what you do and do not offer, your service area, how soon you can book, and what happens next. Then listen to the first twenty calls yourself before you trust it with every lead.

For the same idea applied to the calls that come in rather than go out, see answering service appointment scheduling. When you are ready to try it on your own leads, create an account and hear Milo call you in the browser first.