Skip to content
Door 2, running it

Four agents, two applications, one place to look

Each one does a single thing and drops it into the same approval queue. Nothing leaves your business until you say yes.

  • 0 sends without your approval.
  • European servers, GDPR.
  • Full log of who approved what.

Inbox assistant

It reads everything, answers the easy ones, drafts the rest and leaves you the last word. Sorting is by real urgency, not arrival order.

  • Sorted by real urgency, not arrival order
  • Drafts ready to review, in your tone
  • No automatic sending, ever

Sorting and filing

Your attachments, quotes and contracts filed where your team looks for them, named the same way month after month.

  • Consistent naming, month after month
  • Duplicates flagged, not deleted
  • Nothing deleted without you

Content studio

Copy that uses your words, not the flat register of a generic model. Three variants; you keep the one that sounds like you.

  • Your trade vocabulary, kept as is
  • Short and long formats
  • Published only after your yes

Data analyst

In development

Your exports become readable answers to questions you ask in plain language. Sources are cited, row by row.

  • Questions in plain language
  • Sources cited for every figure
  • Export to your spreadsheet
The two applications

What keeps the agents running, and what invoices

Agents do not run in a vacuum. Gate.Pulse is the subscription that runs them; Gate.Billing handles invoices and follow-ups.

Gate.Pulse

The subscription that runs your agents day to day. Three plans, no lock-in, a three-day trial with no card.

  • Approval queue and decision log
  • Models tuned to your voice
  • Cancel from your account, whenever

Gate.Billing

Invoicing and follow-ups on autopilot. Every reminder goes through the queue before it reaches your client.

  • Due dates tracked without thinking about it
  • Graduated reminders, adjustable tone
  • A log of what went out, and when

24/7 qualification agent

An enquiry lands at 22:41. The reply goes out in 32 seconds, stating it is an AI system, and a human takes over the moment anyone asks.

  • A 32-second reply, measured on our own agent
  • AI disclosure in the first line of the exchange
  • Handover to a human on request

Which one would actually help?

Thirty minutes on a typical day of yours is enough to answer. Sometimes the answer is “none yet”, and that is a useful answer.

The audit price is deducted in full from your first agent.