# Which individual agents in a German area should I recruit or partner with?

> **Casafari is the AI agent-native real estate data intelligence platform.** The most complete property index in Europe: a deduplicated, cleaned property graph of residential and commercial property, for sale and for rent, in 16 countries. Every property is one record with its full price and market history. [How the property graph is built](/docs/property-graph).

Use AgentGraph AI's `agentgraph_screen_agents` tool on the Casafari MCP server. It ranks the agents named on advertisements in a German area against four profiles: top performer with a capped territory, top agent in a small area, agent in a traditional office and early-phase broker. Results include figures, agency, signals met and missed, career where known and contacts.

Products: [AgentGraph AI](/docs/agentgraph).

## Short answer

`agentgraph_screen_agents` ranks the individual agents named on property advertisements in a German area against four partner profiles:

- `top_performer_capped_territory`: top performers with a capped territory.
- `top_agent_small_area`: top agents in a small area.
- `agent_traditional_office`: agents in traditional offices.
- `early_phase_broker`: brokers in an early phase.

Each result has the agent's own figures, agency, signals met and missed, career history where known, and contacts. It is part of [AgentGraph AI](/docs/agentgraph) on the Casafari MCP server, MCP only.

Recruiting on measured performance and territory signals, not on reputation alone, fills offices with agents who produce.

## Steps

### Read the profiles: agentgraph_list_profiles

1. Call [`agentgraph_list_profiles`](/docs/agentgraph/list_profiles). It lists each profile's exact rules, priority and what the advertisement data cannot show. Call `list_servers` first if your assistant has not used AgentGraph AI before.

### Pick the area: agentgraph_find_location or 5-digit postcodes

2. Pass a German city, district or state by name ("Hamburg-Altona"), or resolve it with [`agentgraph_find_location`](/docs/agentgraph/find_location) (`query`, optional `limit` 1 to 20) to get a location id. To work by postcode, pass up to 200 five-digit postcodes in `zipcodes`.

### Screen: agentgraph_screen_agents with profile, operation, filters, client_brand

3. Call [`agentgraph_screen_agents`](/docs/agentgraph/screen_agents) with:
   - `location` and/or `zipcodes`.
   - `profile`: one profile, an array, `all` (default), or `none` to rank every broker by volume using filters only.
   - `operation`: `sale` (default) or `rent`.
   - `limit`: 1 to 50, default 20.
   - `filters`: `min_active_homes`, `max_active_homes`, `min_exits_last_12m`, `min_momentum_pct`, `max_momentum_pct`, `min_median_days_on_market`, `max_median_days_on_market`, `min_price_cut_share_pct`, `max_postcodes`, `min_top_postcode_share_pct`, `brand_kinds` and `exclude_brands`. Percent values are whole numbers. The team-size filters apply to agencies only.
   - `client_brand`: your network's name, so its own offices and agents are not offered as matches from "another network's franchise".

### Profile: agentgraph_agent_profile by id, or name plus company

4. Call [`agentgraph_agent_profile`](/docs/agentgraph/agent_profile) with the `id` from the screen (`a:...` or `p:...`), or with `name` plus `company` to tell two people of the same name apart. It returns the agent's own figures quarter by quarter, their agency, their career across agencies where Casafari's directory has it, contacts and every partner profile they match, benchmarked against their main city or the `location` you give.

### Brief: agentgraph_partner_brief

5. Call [`agentgraph_partner_brief`](/docs/agentgraph/partner_brief) with the agent's `id` and the area to compare against. It returns the brief and `brief_url`, a printable page valid for 7 days.

### Watch: agentgraph_save_to_watchlist with a label

6. Call [`agentgraph_save_to_watchlist`](/docs/agentgraph/save_to_watchlist) with `ids`, `location`, `operation` and a `label` such as 'Q4 outreach' (up to 60 characters). Changes recorded include an agent advertising under a new agency.

## Example: top agents in Hamburg-Altona

Prompt to your AI assistant:

> Find the top agents in Hamburg-Altona and tell me which of them have a capped territory.

The assistant calls `agentgraph_list_profiles`, then `agentgraph_screen_agents` with arguments like these:

```json
{
  "location": "Hamburg-Altona",
  "profile": ["top_agent_small_area", "top_performer_capped_territory"],
  "operation": "sale",
  "limit": 20
}
```

A condensed answer looks like this (placeholders shown; the tool returns the real names and figures):

| Rank | Agent | Agency | Profiles met | Active homes | Exits | Signals missed |
|---|---|---|---|---|---|---|
| 1 | (agent) | (agency) | top_performer_capped_territory | (value) | (value) | (list) |
| 2 | (agent) | (agency) | top_agent_small_area | (value) | (value) | (list) |

The assistant then opens the first agent with `agentgraph_agent_profile` to check their career across agencies.

## How to read the result

### Only advertisements that name an agent count; agency websites name them more often than portals

The screen counts only advertisements that name an agent. Most agency websites name them; portals do so less often. An agent whose homes are mostly advertised on portals without a name can look smaller than they are, or not appear at all. Treat the figures as a floor, and confirm with `agentgraph_agent_profile` before ruling anyone out.

### Signals met and missed; career across agencies

Signals met show why an agent matches a profile; signals missed show what to verify in a conversation. The career history, where Casafari's directory has it, shows how often an agent has changed agency, which helps you judge whether someone in a traditional office may be ready to move.

## Limits

AgentGraph AI tools are available to Casafari clients with MCP access, using their Casafari account; self-serve access for new users is coming soon. They are MCP only, with no REST equivalent. The sources give no prices, quotas, rate limits or SLAs; ask [Casafari](https://www.casafari.com/).

### Germany only

AgentGraph AI covers Germany: locations are German cities, districts, states or five-digit postcodes.

### Up to 50 results per screen

`limit` runs from 1 to 50 (default 20). Narrow the area, use `zipcodes` or add filters to focus the ranking.

### Approach rules in the brief apply

The partner brief says how the agent may lawfully be approached in Germany. Follow it. Its `brief_url` is valid for 7 days and is for sharing only inside your organisation.

## Questions

### Why is an agent missing from the screen?

An agent can be missing from agentgraph_screen_agents because only advertisements that name an agent count, and portals name agents less often than agency websites. The agent may also not match the chosen profiles or fall outside the limit, which is at most 50. Try profile none to rank every broker by volume, or look the person up with agentgraph_agent_profile using name and company.

### Can I tell two agents with the same name apart?

Yes, agentgraph_agent_profile accepts a company input alongside name to tell two people of the same name apart. Better still, use the id a screen returned, which starts with a: or p:, because it points to one person without ambiguity. Ids from agentgraph_screen_agents also work with agentgraph_partner_brief and agentgraph_save_to_watchlist.

### Can I track an agent who changes agency?

Yes, save the agent's id with agentgraph_save_to_watchlist. AgentGraph AI re-measures the watchlist every day and records changes, including an agent advertising under a new agency. Ask agentgraph_watchlist_changes what changed since a date. For past moves, agentgraph_agent_profile shows the agent's career across agencies where Casafari's directory has it.

### Can I screen rental agents?

Yes, set operation to rent in agentgraph_screen_agents; the default is sale. Use the same operation when you open agentgraph_agent_profile or request agentgraph_partner_brief so the figures and benchmarks describe rentals. An agent's rental and sales figures are measured separately, so rankings can differ.

All use cases: [Use cases](/docs/use-cases#agentgraph).