Automation Studio — Israel

Your routine work, running without you.

We design and build automation systems on n8n, Make and AI — from the first working flow to a setup your team stops thinking about.

First call is 40 minutes. You leave with a written scope, free.

Workflow diagram: leads from a website form and WhatsApp are scored by an AI step, then a CRM deal is created, a manager is alerted and a follow-up is scheduled. Website form WhatsApp AI qualifier — reads & scores CRM deal created Manager alert Follow-up scheduled

Diagram of a delivered project — lead intake for a client running two sales channels.

WE BUILD ON TOOLS YOU ALREADY PAY FOR

Automation Core

  • n8n
  • Make
  • Zapier
  • Webhooks & APIs

AI

  • OpenAI
  • Google Gemini
  • Claude
  • RAG on your docs

Data

  • Airtable
  • Supabase
  • Google Sheets
  • Looker Studio

Channels

  • Telegram
  • WhatsApp Business
  • HubSpot
  • Pipedrive

WHAT WE BUILD

Six kinds of work we take off your team

Each one starts as a single flow you can watch run, then grows from there.

Leads and CRM

Every enquiry gets captured, scored and turned into a deal, with the right manager notified before the lead cools down.

n8n · HubSpot · Pipedrive · Airtable · Telegram

Content and publishing

Drafts get written, sent for approval, scheduled and posted — you review, the system does the rest.

n8n · Make · Instagram API · Google Sheets

Documents and reports

Invoices, contracts and weekly reports generate themselves, go out for signature and land where they belong.

Google Docs · SignNow · n8n

Customer support

A bot answers from your own knowledge base around the clock and hands over to a person the moment it should.

Telegram · WhatsApp Business API · Zendesk · RAG

Hiring and onboarding

Applications sort themselves, new hires get their accounts, tasks and first-week schedule without anyone chasing IT.

Notion · Google Calendar · Make

Data and dashboards

Numbers are collected, cleaned and merged on a schedule, so the dashboard is right when you open it.

Supabase · Looker Studio · Airtable

From first call to a system that runs itself

Six stages. You see working software at stage four, not at the end.

  1. 01 Week 1

    Discovery and analysis

    We map how the work actually runs today and find the two or three places where automation pays back fastest. You get the map whether or not we continue.

  2. 02 Week 1–2

    Plan and scope

    We write the spec: which systems connect, what each step does, and the numbers we will judge the result by. Nothing gets built before you have signed off on it.

  3. 03 Week 2

    Architecture and design

    We design the flows, decide who approves what, and define how the AI steps behave — including what they are not allowed to do on their own.

  4. 04 Week 3–4

    MVP sprint

    We build and launch one real process end to end. You watch it run on your own data, and it stays running from that day.

  5. 05 Week 4–5

    Testing and refinement

    We run every branch against real cases, fix what breaks, and keep tuning until the flow survives the messy inputs your team actually sends.

  6. 06 Ongoing

    Support and evolution

    We keep it alive, monitor failures, and add the next process when you are ready. Handover documentation comes with it.

WHAT IT SAVES

Put your own numbers in

A rough estimate, not a quote. We show a range because no process automates to a hundred percent — some of it still needs a person.

Hours returned per month

132 192  h

Money returned per month

$3,300 $4,800

Money returned per year

$39,600 $57,600

Range assumes 55–80% of the routine is actually automated. Month = 4.33 weeks.
Your real figure depends on how messy the inputs are — that is what the first call is for.

4 processes live in production right now
3 weeks from first call to a working flow
11 months the longest-running client flow has been up

WHAT WE BUILT

Four systems, still running

Client names are under NDA. Everything else — the industry, the timeline and the numbers — is exactly as it happened.

E-COMMERCE · 6 PEOPLE · 3 WEEKS

Lead intake across two channels

Enquiries from the website form and WhatsApp land in one place, get scored by an AI step, and become a CRM deal with the right manager already notified.

before: 4 h after: 2 min

average first response time

n8n · HubSpot · WhatsApp Business API · OpenAI

B2B SERVICES · 12 PEOPLE · 2 WEEKS

Content pipeline with approval

Posts get drafted, sent to a reviewer in Telegram, and published on schedule once approved. Nothing goes out unreviewed.

before: 9 h after: 1 h

weekly hours on publishing

n8n · Telegram · Google Sheets · OpenAI

CLINIC · 4 PEOPLE · 4 WEEKS

Support bot on the clinic's own documents

Patients ask in WhatsApp and get answers drawn from the clinic's own price list and protocols. Anything clinical is handed to a person immediately.

before: 62% after: 18%

share of questions reaching staff

WhatsApp Business API · RAG · Supabase

LOGISTICS · 9 PEOPLE · 5 WEEKS

Reporting that assembles itself

Shipment data from three systems is collected nightly, reconciled, and pushed into one dashboard that is correct when you open it in the morning.

before: 6 h after: 0 h

manual hours spent per week on reports

n8n · Airtable · Looker Studio