PLANNEDYou clicked /docs/modules/website-analyzer. This article is part of the planned documentation set — the template below shows the shape every article will take.
Docs/Modules/AI Sales Agent
Modules 9 min read

AI Sales Agent

The autonomous prospector and outreach worker. Reads the market, builds prospect lists, drafts personal outreach, and hands warm replies off to your team.

Overview#

AI Sales Agent is FullView's autonomous outreach worker. It watches the market for prospects that match your business, reads their public signals, drafts personal messages, and sends them at a cadence you approve. When a reply lands, the agent scores it, extracts context, and hands the warm ones off to your team in Conversations.

Think of the agent as the layer between who you're going after and the actual conversation. It doesn't replace judgment — every send is sequenced according to rules you set, and every reply is a human decision from there.

Illustrative
The screens, prospect scores, and outreach examples used throughout this article are placeholders. The shipped article will pull real product UI and real workflow terminology from the AI Sales Agent module as it stabilizes.

How it works#

The agent runs a continuous loop across four stages. Every stage produces data the next one consumes — this is why the agent gets sharper the longer it runs.

01
Read
Watches market signals and your existing prospect data.
02
Match
Scores prospects against your ideal-customer profile.
03
Draft
Writes messages tuned to each prospect's context.
04
Handoff
Routes warm replies into Conversations for the team.
The agent's continuous loop. Every stage feeds the next; every cycle refines the next one.

What the agent reads and produces#

ReadsProduces
Business context (from your workspace)Prospect lists scored against your ICP
Public prospect signalsPersonalized outreach drafts
Reply history and outcomesReply classifications (warm / cold / not now)
Your team's manual editsA continuously refined tone model
Tip
Let the agent learn your voice
The agent adapts to how you actually write. In the first few weeks it's worth reviewing drafts and editing rather than rewriting from scratch — every edit teaches it what you'd have said instead.

Setup#

Getting the agent live takes four steps. None of them require code, though a technical integration path is available in the API Reference for teams routing outbound from an existing CRM.

  1. 01
    Describe your ideal customer
    Point the agent at the type of business you sell into. This is the ICP; it becomes the scoring surface for every prospect the agent evaluates.
  2. 02
    Connect your outbound channel
    Authorize the mailbox or sending domain the agent will use. See Connectors → Connecting your systems.
  3. 03
    Set sending rules and cadence
    How many messages per day, at what times, and with what fallback rhythms. The agent honors these limits exactly.
  4. 04
    Approve the first batch
    The agent generates the first outreach cycle for human review before it sends anything. Once approved, it starts running.
ILLUSTRATIVEProduct screenshot placeholderAI Sales Agent workspace
123
A placeholder for the AI Sales Agent dashboard. The final article will surface an annotated capture of the live UI.

Working with the agent#

Once the agent is live, most of your interaction happens in two places: the agent dashboard (for status, throughput and rules) and Conversations (for the replies it hands off). The dashboard's daily view shows what went out, what came back, and what the agent wants to try next.

During onboarding the agent asks for approval on every send. Once you've approved a few dozen, you can lift approval on high-confidence drafts and keep it on borderline ones — the agent tags each draft with a confidence score.

Warning
Don't fabricate limits
Any specific numbers used in this article (batch sizes, confidence thresholds, opt-out timing) are illustrative. The real defaults and configurable ranges will be documented when the module ships.

Campaign handoff#

When the agent classifies a reply as warm, three things happen at once: the thread moves into Conversations, the prospect is tagged in the workspace's shared context, and — if the reply matches a running Outreach Campaign — it's linked back to that campaign's dashboard. This is the connection point between AI Sales Agent and the rest of the Engagement layer.

  • Warm replies land in Conversations → Inbox with the agent's context notes attached.
  • The prospect's record is updated in shared business context so every other module sees them.
  • Any linked Outreach Campaign shows the reply in-line, so the campaign story stays complete.

Common questions#

Yes. The agent supports sending through your existing outbound channel — see Connectors for the specific integrations. The agent never writes into a mailbox it doesn't own the send credential for.
Every workspace has a suppression list. Add a domain or a specific prospect and the agent will never touch them, retroactively and going forward.
It honors the rules you set. Cadences are configured per timezone; the agent will hold sends until the window opens, even if it has drafts ready.
The agent re-scores its prospect pool against the new ICP on the next cycle. Existing threads keep running; new sends use the updated scoring.
Feedback
Was this page helpful?
← PreviousModules overviewNext →Website Analyzer
Related
Outreach CampaignsDocumentation Working with recommendationsDocumentation Connecting your systemsDocumentation API ReferenceAPI Reference