How to add prospect search to an AI SDR agent with MCP

Map of Europe. Share of companies that publish an e-mail address, where this is known.The countries are listed in the table under the map.Germany: 67%Spain: 70%France: 52%United Kingdom: 62%Italy: 69%Ukraine
Share of companies that publish an e-mail address, where this is known.

Lightest 52%, darkest 70%. The table has the exact figures.

CountryShare
Spain70%
Italy69%
Germany67%
United Kingdom62%
France52%
  1. Agent sends a natural-language querye.g. "Find independent menswear retailers in Germany and Austria"
  2. Prosp matches companies to the queryUses industries, locations, sizes, keywords and exclusions
  3. Contacts found on each company websiteAlso pulls from other public web pages and open registers
  4. Agent receives companies with contactsName, website, industry, size, phones, e-mails, social profiles
The agent describes target companies, Prosp matches them and returns each one with the contacts it publishes.

What Prosp does in the agent

Prosp is the prospecting step. Your agent describes the companies it wants in one natural-language sentence and gets back companies that were matched against that description, each with contacts found on its own website and elsewhere on the public web.

Sequencing, sending and handling replies stay in your own agent. Prosp fills the top of the pipeline with companies and the contact details those companies publish.

What comes back for each company

Each returned company carries its name, website domain, industry, size, headquarters address, the phones and e-mail addresses it publishes, its social profiles and, where a register is open, its registration number.

Each contact carries the person's name, position, e-mail and phone when the company publishes them. You can see where the data originates: the company's own website, other public web pages, and open company registers.

How the agent describes what it wants

The find_prospects tool takes a natural-language query plus optional industries, locations, company sizes, keywords, exclusions, sub-country regions and business models.

The REST API starts the same search with POST /api/v1/discoveries, by a saved ICP id or inline criteria. Credits, rate limits and the returned fields are the same on MCP, REST and Telegram.

Example request and response

An agent sends: Find independent menswear retailers in Germany and Austria.

Prosp returns companies whose profile matches that description. Each one includes the fields listed above—website, industry, size, published e-mail addresses, phones, social profiles and registration number where an open register provides it.

How much contact data is available per country

Coverage varies by country. The map below shows the share of companies that publish an e-mail address, where this is known.

Phone coverage is higher in some markets: Poland and Switzerland both reach 83%, Belgium sits at 82% and Czechia at 80%. LinkedIn pages are the least common signal; Norway leads at 39% and Finland follows at 38%.

Connect your agent

Connect your agent over MCP using the button below. Once connected, the agent can send its first find_prospects request and get companies back with their published contacts.

Questions

Does Prosp send emails or manage sequences?

No. Prosp returns companies and their published contacts. Sequencing, sending and reply handling stay in your own agent.

Can I use the same search from Telegram?

Yes. The Telegram bot runs the same discovery job and returns the same fields. The cost is one cent per search regardless of route.

What costs one discovery job?

One discovery job costs 100 credits, the same on MCP, REST and Telegram.

Connect your agent