INTERCUT case study: an AI sales agent for the German B2B market

Cold outreach without the manual grind

Cold outreach usually means hours of repetitive work: finding companies, browsing websites, saving contacts, guessing relevance and writing messages that do not sound like spam.

For INTERCUT, we built that process as a sales system. The goal was not just to find an email address. The goal was to understand the company, classify its role and prepare a useful sales message in business German.

The brief: find partners and understand their role

INTERCUT needed a systematic way to discover B2B partners in Germany. Some companies were potential production customers, others were better suited for distributor cooperation.

A plain contact list was not enough. The system had to read public company context, prioritize contacts and create a message that matched the business situation.

How the system works under the hood

Leeeds combines public company data, deep website reading, AI classification and human approval. The output is not just a spreadsheet of contacts. It is a prepared sales action.

Contact hunter

The system searches public sources by region and segment, turning random manual research into a repeatable pipeline of companies, websites, phone numbers and commercial signals.

Deep website research

When an email is not visible immediately, the crawler checks contact pages, team sections and legal information. Noise is filtered out and higher-value addresses are prioritized.

AI classification and message angle

AI reads the available website content and classifies whether the company looks like a manufacturer, buyer or distributor. The sales angle changes accordingly.

Human final approval

Leeeds prepares a draft instead of blindly sending. A salesperson can review the company analysis, edit the message, approve it or skip it.

Why this is more than a scraper

A scraper can produce rows in a spreadsheet. Leeeds goes further: it combines contact data with business context, safety checks, campaigns and a client-facing review layer.

The client does not receive just an export. They get a working flow: new contacts, company context, message drafts, pre-send review, campaign history and next actions after replies.

How the message changes by company type

Manufacturer: "I saw that you build CNC machinery and work with assembly processes. It may be worth discussing industrial adhesives for faster and cleaner production."

Distributor: "I noticed you supply technical products for automotive and industry. We would like to discuss whether our portfolio could extend your offer."

What it delivered

Time

Minutes instead of hours

Research stopped being a manual chore. Salespeople worked with prepared candidates.

Quality

Real company context

Emails used public facts about the company instead of a universal template.

Control

AI with a human brake

Automation accelerated the work while the final decision stayed with a human.

This is not another lead list

The biggest value was not simply finding an email. It was giving the salesperson a reason to contact that company and a first message they could actually use.

That is the principle behind Leeeds: less copying data, more qualified sales actions.

Want a similar AI sales system?

Tell us about your market, customer or product. We will show you how contact discovery, qualification and the first outreach can work in your niche.

Discuss the solution

This case study describes a concrete Leeeds sales workflow for INTERCUT.