Connect the company to Bonzo
An administrator adds the company connection. BrokerBot validates the credential, discovers the company context, and prepares pipeline and stage synchronization.
From Bonzo connection to live assistant
Start with boundaries. Prove the behavior. Then turn it on.
BrokerBot’s setup is designed to make each dependency visible: connection, identity, schedule, model, pipelines, stages, tone, and a verified test before production.
Admin
Link each loan officer
Each officer can connect their Bonzo seat and configure their name, NMLS, Calendly, Loanzify, monthly lead target, and reply head-start window.
Define scheduling boundaries
Set the booking window and blackout dates. These settings guide appointment proposals and prevent the assistant from offering times the team cannot support.
Review the plans and configured AI model
Essential is $50/month + $0.01 per BrokerBot-sent message, Pro is $150/month + $0.02/message, and Max is $400/month + $0.05/message. AI processing is additional. The setup view compares the models without changing billing; an administrator requests the commercial tier through BrokerBot.
Select pipelines and stages
Load the current Bonzo structure and explicitly choose where the assistant may work. Configure proactive, reactive, or pipeline-only behavior per pipeline and per stage.
Give each assistant a voice
Choose a personality, name the assistant, and add officer-specific standing instructions. These additions sit above the managed core rules without disabling them.
Test representative conversations
Use the private testing ground to select an officer, lead state, pipeline, and stage. Run a simulated opener and continue the thread with the same production reply stack.
Turn on deliberately
The assistant remains under company and per-officer switches. Start with a controlled scope, review live activity and conversations, then expand as the operating team gains confidence.
What happens on an inbound message
A reply is the end of a pipeline, not the start of a guess.
- 1The webhook or sync process identifies the company and assigned officer.
- 2The current Bonzo history is imported or refreshed, including messages sent outside BrokerBot.
- 3Stage, pipeline, notes, state, model, scheduling, and officer rules are assembled.
- 4The model drafts a response and structured intent metadata.
- 5Deterministic checks can rewrite, delay, regenerate, or block the draft.
- 6An allowed reply is sent through Bonzo, persisted, and reflected in the next action or stage workflow.
Inbound processingExample
01
ReceiveLead reply detected
02
SynchronizeBonzo history and lead context
03
GenerateSelected model + managed rules
04
ValidateSafety, timing, human activity, duplication
05
DeliverBonzo SMS endpoint + event record
What your team sees
Operate from a dashboard built for answers.
The dashboard separates the daily work of a loan officer from the configuration and audit work of an administrator.
Snapshot and rates
See readiness, high-level activity, and reference mortgage-rate information without leaving the app.
Conversations and live monitor
Browse lead threads or switch to the live event view for sends, inbound replies, pipeline actions, skips, and failures.
Assistant controls
Manage personal settings, add to the bot brain, review the configured model and samples, review learning proposals, and use the private testing ground.
Setup and import
Connect Bonzo, manage officer seats, invite the team, select pipelines, and import lead lists into a chosen destination stage.
Company operating rules
Set stage modes, scheduling windows, blackout dates, automation controls, company defaults, and the master switch.
Usage and billing
Review plan status, message volume, model usage, processing estimates, and billing controls.
Common questions
Before you connect.
Does BrokerBot replace Bonzo?
No. BrokerBot uses Bonzo as the CRM and delivery surface. It adds AI conversation, follow-up logic, stage automation, testing, and operational controls around the Bonzo data your team already uses.
Does every loan officer need a separate paid seat?
No. Each plan is priced per company and includes every loan officer: Essential is $50/month + $0.01 per BrokerBot-sent message, Pro is $150/month + $0.02/message, and Max is $400/month + $0.05/message. AI processing is additional and shown separately in estimates—OpenAI through BrokerBot for Essential and Pro, Anthropic direct for Max.
Can we prevent proactive texting in a sensitive stage?
Yes. Set that stage to Reactive if the bot should only answer inbound texts, Pipeline-only if it should update but never text, or remove the stage from the enabled boundary entirely.
What if an officer answers first?
A configurable reply-delay window gives the officer the first opportunity. Recent human activity is also checked before sends so the assistant does not intentionally talk over the team.
Can we test our actual configuration?
Yes. The signed-in testing ground can use the selected officer’s identity, brain additions, model, pipeline, stage, lead state, and scheduling guidance without sending a real SMS.