# Acquisition agent

{/* AUTO-SYNCED SOURCE: this page lives in apps/app/src/modules/db-leads/docs/ and is mirrored into the docs app by `bun sync:module-docs`. Edit it in the module, not in apps/docs. */}

<Lead>
The acquisition agent is the status engine of DB-Leads: it works through each contact's **activities**, understands in context where the person stands, and sorts the lead into the right pipeline phase on its own. It answers "where in the process is this contact?", while the [database scan](/tools/db-leads/datenbank-scan) answers "how likely is a sales mandate here?". Together they make up owner mining.
</Lead>

## What it does

Instead of just static contact fields, the agent reads the **connected activity history** from onOffice (emails, calls, appointments, portal inquiries, notes) and evaluates them in context:

<FeatureGrid>
  <Feature icon="sparkle" title="Understand activities">
    How recent is the last contact? Who wrote last (inbound/outbound)? Is interest escalating or has it gone quiet?
  </Feature>
  <Feature icon="check" title="Derive status">
    That yields a clear status: New, Engaged, Hot, Scheduled, Negotiating, Dormant, Lost, or Won, each with evidence from the activities.
  </Feature>
  <Feature icon="users" title="Sort automatically">
    The agent assigns the lead to the matching phase of your pipeline: automatically, as a suggestion, or not at all, per your setting.
  </Feature>
  <Feature icon="link" title="Reproducible & AGG-compliant">
    The status is derived deterministically from behaviour (never personal characteristics): same activities, same result.
  </Feature>
</FeatureGrid>

## How it flows

<Steps>
  <Step title="Import activities">
    The onOffice sync pulls each contact's activities alongside the addresses (see [onOffice import](/tools/db-leads/onoffice-import)). It is the same data foundation the database scan uses.
  </Step>
  <Step title="Qualify">
    On the AI briefing (single, inside the [database scan](/tools/db-leads/datenbank-scan), or automatically after a sync) the agent works through the activities and derives the status.
  </Step>
  <Step title="Sort">
    Based on the status the agent recommends a pipeline phase and applies it according to your automation setting. When you accept a finding from the scan, the lead moves into the recommended phase as well.
  </Step>
</Steps>

## Automation setting (per workspace)

Under **Settings → REOS AI** you decide how the derived phase is applied:

<DefinitionList>
  <DefItem term="Automatic">The agent moves leads into the matching phase itself. Every move is logged as an AI activity and is reversible.</DefItem>
  <DefItem term="Suggest">The agent shows status + recommended phase; you accept it with one click on the customer view.</DefItem>
  <DefItem term="Off (default)">Only show the status, never change the phase automatically. Every workspace starts here; you enable the automation deliberately.</DefItem>
</DefinitionList>

<Callout tone="note" title="The human stays in control">
Even in automatic mode the agent only acts during a qualification (manual, inside the scan, or automatically after a sync), never silently in the background. Every move is in the history and can be undone.
</Callout>

## Staying current automatically

A sync fetches data, it does not rate anyone. If you want new and changed contacts assessed without lifting a finger, set the **auto-scan** under **Settings, Database scan** to daily, weekly or monthly. Every automatic run syncs with onOffice first and then rates only the contacts where something happened since the last assessment. Data first, rating second, never the other way round. The scan analyses are included in the DB-Leads add-on.

## Which activities count

Under **Settings → Qualification → Context** you define which activity types feed the assessment. Without rules all activities count; with rules only the matching ones.

## Status at a glance

The derived status is visible everywhere: as a chip in the list and board, prominently on the [customer view](/tools/db-leads/kundenansicht) (with evidence and last activity), and as a dedicated **status filter** in the list, so you can work through, say, all "hot" leads at once.

## Frequently asked questions

How status and seller score relate, who decides about a lead and what works without an AI provider: [Questions and answers](/tools/db-leads/faq).