The challenge

About 400 new sign-ups arrived every month. The sales team, 2 people who also worked as account managers, had to check them one by one to decide which were worth contacting. That was hours of repetitive work taken away from selling and from existing clients, while the best leads waited.

What we did

What people did

With the sales team we defined the ideal customer profile and the signals that separate a good lead from an uninteresting sign-up:

  • Whether the email is a business address or a free one.
  • Whether the company exists.
  • How many domains are in its portfolio.
  • Whether real employees can be found on LinkedIn.
  • Its Trustpilot score.
  • Whether the business conflicts with the terms and conditions, such as gambling or the sale of illegal goods.

What AI does

When someone creates an account, the CRM (HubSpot) sends the data to the agent by webhook, in real time. The agent, built on Claude through the Anthropic API, checks the criteria, gives a score from 1 to 100 and posts it in Slack, with sign-ups ranked and a short explanation. The team still decides who to contact, and books meetings with the best leads.

How the agent works

As soon as the CRM receives a new sign-up, it sends it to the agent by webhook. Within seconds it lands in Slack with a score and the reasons. The sales team decides who to contact.

Role

You are a sales analyst at a domain registrar. Score every new sign-up.

Ideal customer

A real company with staff, several domains and a good online reputation.

Weighted criteria

  • Business email20%
  • Company exists20%
  • Domains in portfolio20%
  • Staff on LinkedIn15%
  • Trustpilot rating10%
  • Conflict with T&Cs15%

Output format

{
  "score": 1-100,
  "fit": true | false,
  "reasons": ["...", "..."]
}

The prompt sets the agent's role, the ideal customer, the weighted criteria and the output format.

The results

  • About 400 sign-ups a month scored automatically, with no manual checks.
  • The team sees straight away in Slack which leads are qualified and which are not.
  • The 2-person team won back many hours for selling and looking after clients.

What this can do for you

If leads or sign-ups arrive every day and someone has to filter them by hand, an AI agent built on your own criteria can triage them instantly and deliver the result where your team already works. Your sales team focuses on the best leads, without hiring an in-house engineering team.

The AI scoring agent gives us total clarity on every sign-up the moment it happens, letting us reach out to top leads in real time. It's like adding a dedicated research team without increasing our headcount.
Ria, Business Development Manager

Services