What to automate first — and how to tell if it's working.

Automation means a computer doing a task a person currently does by hand: sending an email, updating a spreadsheet, logging a lead, booking a meeting.

For most small teams it isn't the important work that piles up — it's the copy-paste, the follow-ups and the data entry around it. That's what automation is for.

We mostly use n8n: a visual workflow builder that connects 400+ apps — Gmail, Slack, Notion, HubSpot, Cal.com, WhatsApp and more — and lets you define exactly what happens, step by step, when something occurs.

1. 01

   Does it happen more than a few times a week?

   A one-off task isn't worth automating. The payoff comes from repetition.

2. 02

   Is it predictable?

   Automation works best when the same input should always lead to the same result. If every case needs judgement, keep a person in the loop.

3. 03

   What happens if it breaks?

   Tasks that move money or delete data need more testing and fallbacks. Notifications and logging are safe to automate quickly.

1. 01

   Lead capture → CRM

   Every form, DM or email enquiry logged automatically, tagged with where it came from. Nothing falls through the cracks.

2. 02

   Meeting booked → briefing

   When a call is booked, an AI pulls together the prospect's website and past emails into a one-page brief, so you walk in prepared.

3. 03

   Brief → proposal draft

   A short form with the client's goal produces a first-draft proposal in your format, ready for you to edit and send.

4. 04

   Invoice → follow-ups

   Reminders go out on a schedule and stop automatically the moment payment lands.

5. 05

   New client → onboarding

   Contract signed: project created, welcome email sent with a checklist, team channel set up, kick-off call scheduled.

A simple way to estimate it:

Example: two hours a day of manual lead handling at ₹1,000/hour is about 43 hours a month — roughly ₹43,000 of time. If the automation costs a fraction of that to build and run, it pays for itself quickly. If it doesn't, it isn't worth building yet.

On the free call we'll do this rough maths with you for the tasks you mention.

1. 01

   Automating a broken process

   Automation makes a process faster, not better. Fix the steps first, then automate them.

2. 02

   Skipping error handling

   What happens when an app is down or rate-limits you? Every real workflow needs retries and a fallback path.

3. 03

   Not watching it after launch

   Workflows break when APIs change or logins expire. Set up alerts so failures are noticed, not discovered weeks later.

4. 04

   Doing too much at once

   Start with one workflow, learn from it, then add the next. Automating everything at once usually means nothing works well.

5. 05

   Ignoring data quality

   Messy data in means messy data everywhere, faster. Sometimes a clean-up comes before the automation.

1. 01

   n8n

   Workflow engine — open source and self-hostable.

2. 02

   OpenRouter + Ollama

   Access to models like Claude, GPT and Llama, or local models that keep data on your machine.

3. 03

   Notion, Cal.com, Slack, Gmail

   Where the work actually happens for most small teams.

4. 04

   Supabase / PostgreSQL

   Structured data when a spreadsheet stops being enough.

5. 05

   Next.js + Vercel

   Websites and lightweight internal tools.

Nothing here locks you in. Everything we build can be handed over and maintained by your own team.