Borrow Brains Automation OS

Multi-tenant AI workflows
with lead intelligence.

BBOS runs your client-facing automations on a HubSpot mirror and scores every contact across segment, decision stage, and perceptual filter — so the right outreach lands at the right moment.

Request accessSign in

v0.1 · onboarding pilot tenants · EU-hosted

What it is

Three layers, one runtime.

Workflow engine

Visual editor on @xyflow/react. Cron, webhook, manual, and event triggers. Approval gates, retries, and a model-pinned LLM node — Opus, Sonnet, Haiku per step.

HubSpot mirror

A bridge, not a replacement. Contacts, deals, and engagements sync via webhook plus reconciliation. Three BBOS-owned properties write back; the rest of HubSpot stays canonical.

the moat

Lead intelligence

Three Pareto segments × five awareness stages × six meta-program filters. Every contact gets a triple, every triple resolves to a next best action.

How it works

Wired up in an afternoon.

  1. 01

    Connect HubSpot

    Paste a Private App token. Five minutes. Contacts, deals, and engagements begin mirroring; webhooks land idempotent.

  2. 02

    Train BBOS on your voice

    Upload five emails and five long-form pieces per channel. Anchored as pgvector(1536) embeddings; few-shot grounded into every compose step.

  3. 03

    Approve, ship, measure

    Drafts pass an anti-AI guardrail before reaching your inbox. Per-flow auto-send is earned, never granted by default.

Lead intelligence

Every contact, classified on three axes.

The engine alone is table-stakes. The intelligence layer is what makes each automation sharper than an operator could configure manually.

Each contact gets a {segment, decision_stage, perceptual_filter} triple. Three Pareto segments per tenant, deterministically assigned via embedding + k-means. Five Schwartz awareness stages, from unaware to most-aware. Six Charvet meta-program filters, extracted from existing written communications — no instrument required.

The triple resolves through a rules table to a workflow and a payload. When no rule matches, an LLM proposes one for human promotion.

contact triple · example

Kwame Abedi · Brent PLC

  ├─ segment       P1  "cautious scaler"
  ├─ stage         solution-aware
  └─ filters       risk-averse · process
                   authority-deferring
                          │
                          ▼
                   next best action
                   →  book 15-min demo
                   →  send brief-cpd-21.pdf

  confidence       0.82

Pricing

Three tiers. Pricing on value, not cost-plus.

Starter

Single workflow, one channel.

$299/ month

$1,500 setup

  • 3 automations
  • 10K contacts
  • 500K LLM tokens / mo
  • Shared subdomain
  • Email support
Request access

Pro

Recommended

Production tier. Custom domain, full intelligence layer.

$749/ month

$3,000 setup

  • 10 automations
  • 100K contacts
  • 5M LLM tokens / mo
  • Custom domain
  • Priority email + monthly call
  • 1 year run history
Request access

Enterprise

SSO, SLA, dedicated support.

$1,999/ month

$7,500 setup

  • Unlimited automations
  • 500K contacts
  • 25M LLM tokens / mo
  • SSO
  • Dedicated Slack + weekly call
  • 99.5% SLA · unlimited run history
Request access

All tiers include SSL, custom domain available, SSO available on Enterprise.

Ready to talk?

We're onboarding pilot tenants now. Tell us about your stack and the automations you'd run first — contact@borrowbrains.com.

Request access