Neeraj Shekhawat

Portfolio

One revenue engine, built layer by layer.

Each build covers one layer of how revenue moves: create pipeline, capture demand, manage the pipeline, run the team, plan the business. Status is shown plainly, and nothing here claims a result it has not produced.

The build plan

One revenue engine, five layers

Every build maps to one layer of how revenue moves through a company. Three layers have working systems today.

  1. 01Create pipeline

    Signal-based outbound

    Clay outbound for a market with no data, an n8n outreach workflow and an AI lead bot.

    Live
  2. 02Capture demand

    Inbound Router

    Demo request to the right rep in five minutes, with an SLA watch.

    Partly tested
  3. 03Manage pipeline

    Revenue Desk

    Call briefs, notes to CRM, news risk and a Monday forecast.

    Tested
  4. 04Run the team

    Field Sales Command Center

    Voice-to-CRM visit logging for field reps.

    Planned
  5. 05Plan the business

    Revenue Planning Model

    Capacity, quotas, hiring and unit economics.

    Planned
Live in use Tested built and tested Planned next on the bench See every build

Case studies

Each one covers the problem, what I built, what broke and what I changed, and the trade-off.

Layer 3 · Manage pipeline · Oct 2026

Revenue Desk: a deal desk that runs itself

Tested end to end

Four n8n workflows on one CRM. Briefs before every call, call notes turned into CRM updates, weekly news risk on every open deal, and a Monday pipeline review where code computes the forecast and Claude writes the narrative. It caught $105,000 of commit sitting on two accounts with a new CFO or company-wide layoffs.

n8n Claude API Notion Gmail
Read the case studyWatch on LoomDownload the workflows
From the 11 Oct test run
Commit vs target$198,000 / $300,000
Weighted pipeline$215,300
Coverage1.74x
News flags after the fix10 of 12 → 4 of 12
Layer 2 · Capture demand · Oct 2026

Inbound Router: demo request to the right rep in five minutes

Routing and handoff tested

Reads the lead's website, scores fit in code, routes to the rep with the fewest open leads and drafts the first reply. Qualified leads become deals in Revenue Desk without anyone retyping them. SLA escalation and the weekly report are built and still to be tested end to end.

n8n Claude API Notion Gmail
Read the case studyDownload the workflows
How it decides
Fit scoreOut of 100, in code
Tier A / Tier B SLA5 min / 60 min
RoutingFewest open leads
Qualified 21-50 rep team$35,000 Discovery deal
Layer 1 · Create pipeline · Oct 2026

Clay outbound for a market with no data

Built, not sent yet

An outbound system for small HVAC companies, a buyer with no LinkedIn page and no findable work email. Keyword filtering, hiring signals, a research agent that refuses to guess, and copy built only from verified facts. My cohort coach used its copywriting agent as the class example in a live session.

Clay Claygent Clay Sequencer
Read the case studyWatch on Loom
From the Clay table
Companies after filter278 → 136
Verified work emails4 people
Sequence3 steps, one-word reply
Clay spam checkGreen
Layer 1 · Create pipeline · Oct 2026

An outbound system built to survive re-runs

Certification in progress

Built for Clay's Outbound Certification, which scores reliability rather than copy: re-runs that never double-contact, exclusions that hold, and rows that recover when a data provider fails.

Clay
Read the write-up
What it was graded on
Re-runsNo duplicate contacts
ExclusionsLists that hold
FailuresRows recover, run continues
CopyNot scored

Also built, in the workshop

Built and running in my own n8n. Walkthroughs are on the way.

Local-business outbound engine

Reads each company's website for the owner's name and email, keeps an owner only when the page proves it, then builds a daily call list, syncs call outcomes and drafts a 3-touch email sequence for review.

n8n Claude API Google Sheets
First live run in progress

Job Search Signal Engine

Scores remote GTM and RevOps roles, researches each company and drafts outreach for me to review. Nothing sends on its own.

n8n Claude API Google Sheets Notion
Final step in testing

n8n outreach workflow

Source, enrich, personalize with an LLM draft that a person approves, sequence on days 1, 3 and 7, log to the CRM. Cut manual effort by about 60%.

n8n OpenAI GPT
LiveEarlier build

AI lead-generation bot

Scrapes and scores prospects with custom logic. Only qualified leads get a GPT-written outreach step. The lesson: routing logic matters more than the model.

OpenAI GPT
LiveEarlier build

Earlier builds are documented in my Notion portfolio.

Next on the bench

Layer 4 · Run the team

Field Sales Command Center

Voice-to-CRM visit logging for a field sales team, so a rep logs a school visit from the car instead of at the end of the week.

Planned
Layer 5 · Plan the business

Founder's Revenue Planning Model

A template for capacity, quotas, hiring and unit economics, based on my work on a startup's 5-year revenue model. Built with made-up numbers.

Planned

Hiring for GTM engineering or RevOps? Need a revenue system built?

Email is fastest. I work evenings India time, which overlaps the US morning.

Get in touchLinkedIn