Inbound lead qualification: the agent prepares, the rep decides

It can do what delays the first reply: read the request, retrieve what the company already knows about the account, complete the CRM record and propose a reasoned score. The rep approves the reply and sends it; the agent never contacts a prospect on its own. The split comes down to one simple rule: anything that gets read and prepared can go to the agent; anything that goes out to a prospect goes through a person. Here is what that looks like day to day, and how to check, numbers in hand, that the first-reply delay has really dropped.

Why a lead loses value within an hour

A form filled in on the website, a demo request, a question sent by email: at the moment the person writes, they are available, focused on their need, and often comparing several suppliers. An hour later they are in a meeting or have moved on. The Lead Response Management study, led by Professor James Oldroyd on more than 15,000 leads and 100,000 call attempts, measured that effect: the odds of qualifying a prospect are divided by 21 when the response delay goes from five to thirty minutes, and the contact rate drops more than tenfold after the first hour.

Nobody decides to answer late. The lead comes in at 10:12, the rep is in a meeting until noon; when they get back, the CRM record is empty or outdated and has to be rebuilt before writing; and the request itself deserves a proper read, to know whether it calls for a phone call, an email or nothing at all. It is 3 pm when the reply goes out, and it is nobody’s fault. The authors of the Harvard Business Review article that accompanies the study put it plainly: “Our research shows that most companies are not responding nearly fast enough.”

And sales teams have not waited: according to Eurostat, in 2025, 34.70 % of European Union enterprises that use AI use it for marketing or sales, ahead of every other purpose. So the real question is what you hand the agent, and what you keep.

What the agent reads before answering

A qualification agent does not start from scratch. As soon as a request arrives, it reads the form or the email, then searches the sources the team has opened to it: the CRM, to know whether the account exists, who has already contacted it and where the opportunity stands; the history of exchanges, to find an old request or a quote sent last year; the account’s public information, website and news, within the limits the team has set. That is the job of Scalt’s account enrichment agent: watch target accounts, enrich records and keep the CRM current.

It reads with the rights of the person it works for, no more. If a rep cannot see a colleague’s accounts, the agent preparing their reply cannot see them either. The rule sounds obvious; it is nevertheless what lets you open the agent to the whole team without creating an internal leak.

The reasoned score

A score with no explanation is useless to a rep: they know neither why a lead is rated 80 nor what to check before calling. The agent therefore proposes a score in which every criterion is visible and tied to its source. The criteria are the team’s own, written in plain language, and they change when the strategy changes.

The result fits in a few lines: a score, the reasons behind it, and two or three points the rep would do well to check by voice. When the team disagrees with a score, and it happens, it corrects the criterion. The next score takes it into account.

The first reply, prepared but never sent alone

With the record completed and the score in hand, the agent drafts a first reply: in the team’s tone, with the account elements it has retrieved, and a proposed next step suited to the stated need. That reply lands in the rep’s Inbox as an action to approve. They reread it, edit it if needed, and send it in one click. If they reject it, the agent does not insist and keeps the reason for next time.

We hold to that rule for everything that leaves the company: an email to a prospect is sent by someone. What the agent removes is the preparation time; the approval takes ten seconds and it stays. Follow-ups follow the same logic: the Follow-ups & pipeline agent spots stalled deals and proposes the right follow-up, ready to go, and the rep presses the button.

What stays with the rep

• The call: voice remains the best way to qualify a need, and the agent does not phone.
• The final qualification: the score is a proposal; the rep decides whether the lead enters the pipeline and at what stage.
• The decision to pursue or drop, and the relationship that then builds with the account.
• The scoring criteria themselves: the team writes them, reviews them every quarter and changes them when the market changes.

Prospect data and traceability

A lead is a person, and their data is personal data; your data protection officer will ask the question, and they will be right to. What reassures them, in practice: the agent reads only the authorised sources, with the user’s rights; personal data is anonymised before any model call; and every request opens a case that keeps track of what was read, proposed, approved or rejected, by whom and when. That case also serves the rep: six months later, they can find why the lead was set aside.

Those cases feed the exportable governance log, which the compliance team can consult without going through the sales team. That is what lets you extend the agent to other teams without renegotiating each time what it is allowed to do.

Measure, before and after

No need for one more dashboard. Three numbers are enough to know whether the agent is useful, and they can be read in the CRM it keeps current:

• The first-reply delay, from the lead’s arrival to the reply approved by the rep, as a median: one reply sent on Monday morning to a Friday-evening lead is enough to skew an average.
• The share of CRM records that are complete at the time of the first call: sector, size, contact, history.
• The qualification rate, meaning the share of inbound leads that become an opportunity, measured by source and by month.

Record them over a quarter before the agent, then over the following quarter. If the delay does not move, the problem is not preparation but the reps’ availability, and no agent will change that.

Frequently asked questions