Platform

Mako is the conversation engine. Your CRM is the system of record.

Built AI-first, multi-channel, multi-tenant. From day one.

SMS
Twilio
Voice
Real-time
Email
Resend
One context
Customer profile · Lead metadata · Conversation history
Architecture

Multi-channel, one context.

SMS, voice, and email all share the same conversation history, customer profile, and lead metadata. When a customer texts in the morning and calls in the afternoon, Mako already knows the context.

Agent identity model

One specialist per intent. Swapped under the hood.

Every inbound message is handled by exactly one specialist — trade-in, finance, service, F&I follow-up, or one of the universal specialists: info, appointment, escalate, translate. A triage agent re-routes every turn. The customer never knows they're talking to different specialists. The conversation history stays whole.

Inbound
"Still got that RAV4? Whats the payment on 84 months?"
Triage classifier → finance specialist
Outbound
"It's still on the lot. On 84 with $2k down you're looking at ~$486/mo OAC. Want me to lock it down with an appointment Tuesday?"
Conversation history stays whole
Bolts onto your CRM

We never replace. We always add.

Mako connects via webhook to DealerSocket, VinSolutions, CDK, Reynolds, Dealertrack, Tekion, and others. Every AI conversation gets written back to the lead's activity log in your CRM. Your team takes over by replying inside their CRM — Mako auto-pauses with a 5-minute cooldown that resets every time staff messages.

See all integrations
The customer data firewall

Local AI handles every moment customer data is touched.

When Mako's cloud AI asks for a phone number, name, or email, the local AI on your MakoNode takes over. It detects the customer data, writes it to your CRM, and sends only a metadata event to Mako Central — never the actual value. Even ambient customer data like names mentioned mid-conversation get tokenized before anything leaves your dealership.

Layer 1: AI semantic scrub.

A proprietary on-device language model catches contextual customer data.

Layer 2: Deterministic redaction.

Rule-based checks catch the structural stuff — phones, postal codes, financial identifiers.

Backstop: Mako Central rejection.

Third check at the cloud layer rejects any inbound carrying customer data that somehow got through.

Read the full architecture
Compliance

CASL, TCPA, OMVIC. Engineered, not bolted on.

CASL silent opt-out
English + French
TCPA quiet hours
9 PM to 8 AM customer-local
Immutable audit log
Defensible in regulatory complaints
Multi-vertical from day one

Vertical is a config dimension, not a code fork.

Mako launches with automotive. Mortgage, insurance, real estate, and solar follow. Each vertical gets its own specialists, compliance rules, terminology, and manager workflow — all configured, not custom-built.