Automattic CRM
AI Agent

AI that works for you, not just with you.

Not a chatbot strapped to a CRM. Native AI across every surface — with the provider of your choice.

OOpenAI
AAnthropic Claude
GGoogle Gemini
OlOllama
The Agent

Describe it. The agent does it.

You write one sentence. The agent plans multi-step work across your CRM, asks permission on risky steps, and logs everything.

  • Multi-step planning and execution
  • Real-time streaming of every step
  • Permission-aware — only acts in the user's scope
  • Full audit log of every action
Example prompt
"Find deals over $50k with no activity in 14 days and draft a check-in email."
Example prompt
"Score this week's signups and tag the top 20% as priority."
Example prompt
"Close tickets resolved over 7 days ago and send a survey."
Example prompt
"Summarize every deal I own that's closing this month."

AI across every surface.

Lead scoring

Know who's worth the first call.

Every contact gets a 0–100 score based on engagement signals, with human-readable factors.

Priya Shah · Lumen Labs
Score 87
  • +24 Opened 5 of last 6 emails
  • +18 Booked demo within 2 days
  • +15 Company size matches ICP
  • −6 Hasn't visited pricing page
Q2 forecast
Optimistic
$418k
Realistic
$312k
Pessimistic
$204k
Deal forecasting

Know where the quarter is heading.

Three scenarios, updated daily, grounded in win probability and historical close rates.

Ticket triage

Every ticket, pre-sorted.

Auto-categorize, set priority from sentiment, and draft a reply in your tone of voice.

#8241Export failing on large listsHigh
#8239Custom fields on deals?Normal
#8236Invoice PDF formattingLow
Your AI, your choice

No lock-in. Four providers. One switch.

Pick a provider in settings. Every AI feature re-routes instantly. Bring your own key to skip our credits.

O
OpenAI

GPT-family models.

A
Anthropic

Claude models.

G
Google

Gemini models.

Ol
Ollama

Fully self-hosted.

Privacy & control

Your data never trains public models.

AI calls run scoped to your workspace, with every action recorded in a tamper-evident log you can export any time.

  • Permission-aware actions
    The agent only acts within the invoking user's scope.
  • Complete audit log
    Every AI-initiated action with prompt, model, and outcome.
  • No training on your data
    Zero-retention inference with every supported provider.

Delegate the work. Keep the craft.

Try the agent on your own data — 14 days, no card.

Skip to main content