Skip to content
All use cases

Automation

Lead classification with reasoning

A readable priority list per lead, with arguments, not a black-box score.

A lead form on the site, a whitepaper download, a conference badge scan. At the end of the week eighty names sit in the CRM. Which five should sales call today, and why?

The situation

A B2B software firm with a small sales team gets leads through three channels: a contact form, a whitepaper download, and the badge scans at trade fairs. The CRM (HubSpot or Pipedrive, sometimes even Notion) fills up with names, mail addresses, and an "interest" field that nobody fills in consistently.

The salesperson would prefer to call the leads that turn out to be serious, but figuring that out is itself twenty minutes of work per lead. LinkedIn check, looking at the company site, guessing whether this is serious or a student.

A classic lead score (points per action, mail open, download, visit) says something, but not why. A score of 73 is not an argument to pick up the phone.

What we did

We do not build a scoring platform. We build a daily shortlist.

For every inbound lead the system pulls publicly available information: company size from the chamber of commerce, sector, location, and where relevant a note on their own site about the software stack involved. No scraping of gated content, only what is openly findable.

Then a short description per lead: what kind of company is this, what are they probably looking for, which of our products fits, what would be a reasonable opening question in a first call. Not as a sales pitch, as a briefing for the salesperson.

Sales gets a list of five to ten leads each morning, in priority order, with reasoning. They can override the order, the system remembers that for the next day. No score without explanation, but a readable rationale.

What it delivered

Time from "lead in" to "first human touch" drops from days to hours. Not because the system calls, but because the salesperson no longer has to research before calling.

Also happens: poor leads are visibly recognised instead of guessed. A student downloading a whitepaper for a thesis gets a short friendly mail, not a phone call. That saves time on both sides.

A week in, the salesperson sees which leads actually converted and which arguments sat under them. That sharpens their own judgement for the leads where the system itself is uncertain.

What this wasn't

Not autopilot for sales. A readable priority list with reasoning per lead, not a score without context.