> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lumisreach.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Three agents, one pipeline
> Voice, chat, and email agents that share one contact record, one qualification model, and one hand-off path, running on a live data layer that sources accounts and keeps every record current.
When you deploy LumisReach, you turn on three GTM agents. Voice. Chat. Email. They share one contact record, one set of qualification criteria, and team calendars to book against. They also share a live data layer that sources net-new accounts against your ICP and keeps every record current, so every conversation starts from a complete picture. They hand off to each other, and to a human, without anyone re-typing context.
## The three agents
### Voice agent
Picks up in half a ring. Qualifies live, handles objections in your brand voice, books the meeting before the hang-up. Runs inbound and outbound on the same engine, with Branded Calling on every dial so your business name shows up on every call.
Typical metric: **average answer 0.4 seconds**.
### Chat agent
Stays in the thread, never a ticket queue. Replies in your voice and escalates to a human the instant intent spikes, without the hand-off friction. The transcript continues into the same contact record the voice agent is using.
Typical metric: **first reply 8 seconds**.
### Email agent
Writes the first touch and every follow-up. Personalized at the line level, deliverability tuned, and it keeps firing until the meeting is booked or the contact opts out. Engagement events (opens, clicks, replies) feed back into the voice and chat agents.
Typical metric: **reply lift +312%** on warm-list re-engagement.
### The data layer
The three agents work from one shared record. The data layer keeps that record complete.
**Sourcing.** Define your ICP (industry, size, geography, role, technology) and LumisReach produces a fresh batch of matching accounts and verified contacts on the cadence you pick. Net-new only. Existing CRM rows are excluded. See [Source leads](/solutions/core-solutions/source-leads).
**Enrichment.** Missing fields get filled. Stale rows get refreshed. Unknown inbound callers get identified before the voice agent picks up. Typical metric: **enrichment hit rate \~85%** on US business callers. See [Enrich leads](/solutions/core-solutions/enrich-leads).
## How they work together
A single prospect gets one consistent treatment across channels:
1. The **data layer** sources the account (if it's net-new) and enriches the record before any agent touches it.
2. The **email agent** sends the first touch at the optimal moment based on prior engagement signals.
3. The prospect clicks. The **voice agent** calls within 60 seconds carrying the context of what they engaged with.
4. If the prospect bounces mid-call or prefers text, the **chat agent** picks up on the site without re-asking what they already said.
5. When intent crosses your threshold, the call lands on the right rep's calendar with a structured note, a recording, and a transcript.
One contact record. One qualification model. One booking calendar. Three channels and a data layer working together.
## What a real qualification call sounds like
The voice agent qualifies, books, and writes back the structured fields. The AE wakes up to a meeting on the calendar with the full transcript and the prospect already in the CRM.
> **Agent:** Hi Daniel, I noticed you just downloaded the buyer's guide. I run new-account onboarding here. Have two minutes?
>
> **Daniel:** Sure, for two.
>
> **Agent:** Are you evaluating for your team, or doing research for someone else?
>
> **Daniel:** For my team. We're scoping a decision for next quarter.
>
> **Agent:** Got it. Roughly how many seats are you looking at?
>
> **Daniel:** Around 40.
>
> **Agent:** Based on that size and your timeline, I'd like to put you with our mid-market lead. She has Tuesday at 10am or Wednesday at 2pm open. Which works?
>
> **Daniel:** Tuesday.
>
> **Agent:** Booked. The invite and a one-page on the mid-market plan are on the way.
The same flow runs for legal intake, healthcare bookings, real-estate inquiries, and field-service dispatch. The agents adjust the script to the audio in front of them. The structure (qualify, book, write back) stays the same.
## Boundaries
The agents operate inside well-defined limits.
Answer questions from your knowledge base
Book meetings against the right rep's calendar
Qualify against your scoring criteria
Transfer to a human when intent spikes or the conversation calls for it
Follow up by voice, SMS, email, or another call
Speculate when the knowledge base doesn't have the answer
Promise terms outside your stated policies
Claim to be human when asked directly
Surface or call data we do not have rights to
Handle scenarios you've routed to escalation
When the conversation goes outside the knowledge base, the agent captures contact details, hands off cleanly, and flags the gap in your dashboard so you can teach the team for next time.
## Training the team
You teach all three agents once, in one place. The same knowledge base, qualification rules, ICP definition, and brand voice apply across voice, chat, and email.
Budget, timeline, fit, decision-maker, use case. The agents apply these in conversation, not as a rigid script.
Pricing, product specs, common objections, policies, services. See [Knowledge base](/concepts/knowledge-base).
Hand-off lands in HubSpot, LeadConnector, or Zoho with structured fields populated. Bookings land on the right rep's calendar.
Every conversation produces a transcript and a summary. Gaps surface in the dashboard, and you teach the agents from them.
What goes in your knowledge base
Match the personality to your brand