Skip to content

Subhajit Gorai · GTM engineer · Bengaluru, India

I build AI systems that move GTM metrics. This site is one of them.

Growth product manager by title, engineer by practice. I build and run the enrichment, outbound, and content pipelines that put qualified pipeline in front of sales — and everything below is a real one, with the reasoning attached.

Founder's Office, New Products at Gushwork · Previously Growth PM at Weekday (YC W21) · First product hire at Gushwork. Promoted twice in 12 months.

Here's how I just did that

That headline was picked for you. Here is every signal I checked, in the order I checked it — this is the actual trace from your page load, not a description of one.

Resolving… (if this persists, JavaScript is off — you are seeing the default variant, which is the correct fallback.)

It is a waterfall, and it is the same shape as the enrichment system under Systems: check the most specific signal first, fall through on a miss, always land somewhere valid. The expensive step never runs when a cheap one already answered.

Three signals, all of which your browser hands over as part of a normal page request: two URL parameters and the referring page. No IP lookup, no fingerprinting, no cookies. Nothing here is stored — the only data this site keeps is what you type into a form yourself, and those say so before you type. If a trick can't survive being written out in plain English like this, it doesn't belong on the site.

  • 2xpromoted in 12 months
  • ~75%of annual ARR at a YC company, from the AI outbound engine
  • 100+SEO pages generated in ~5 hours
  • +40%organic traffic from those pages
  • $15KMRR US vertical, scaled from 0
  • 10,000+leads/month through the enrichment pipeline
  • ~85%cut in lead response time
Currently shippingAgent workflows aggregating spend + lead data across 60–100+ Meta ad accounts into a daily tracker with automatic drift flags.
01

Systems

Production pipelines, with the reasoning attached. Select any step in the flagship to see why it is built that way.

Gushwork

AI enrichment pipeline

The problem

10,000+ leads a month for US local businesses, with BDRs doing the research by hand. Speed-to-lead was the conversion bottleneck — not lead volume, not copy, not the pitch. Every hour a record sat unresearched was an hour a competitor could answer first.

The call that made it work

B2B databases like Apollo are weakest exactly where this ICP lives: SMB and local. So the primary source isn't a B2B database at all — it's Google Business Profiles, the authoritative record for local businesses.

Select a step for the reasoning

  1. GBP scraping

    Google Business Profiles are the authoritative database for local businesses. Name, phone, email, website, and reviews come off the profile directly.

    Marginal cost per record is near zero. That single property is what makes everything downstream affordable — you can afford to look at every lead because looking is free.

    This is the inversion: most teams start at a paid B2B database and treat scraping as the fallback. For SMB and local, the paid database is the weaker source. Starting there means paying more for worse data.

leads/month processed
10,000+
cut in lead response time
~85%
Full write-up, build-vs-buy call

Weekday (YC W21)

AI outbound engine

Semantic candidate matching into automated outreach, replacing manual sourcing — drove roughly 75% of annual ARR.

~75%of annual ARR driven
98%match accuracy on skill search
See the system

Gushwork

SEO generation agent

Keyword research through to published page, fully automated — 100+ pages in about five hours.

100+pages generated in ~5 hours
+40%organic traffic
See the system
02

Tools

Free, no signup. Each answers a question I had to answer at work, and each shows its assumptions — so you can disagree with the output instead of trusting it.

Enrichment cost calculator

What a credit-based platform costs versus running the same waterfall yourself — at your volume, with your fields.

10,000
Fields you need
30%

The share your free source misses. This is the number that decides the answer — and the number most teams have never measured.

Credit-based platform (Clay-style)

$949

40,000 credits/mo

Custom pipeline (scraping + direct APIs)

$550

$400 fixed + usage

At 10,000 leads/month, custom is $399 cheaper per month.

$0$2,075$4,149025k50kbreakeven 4kleads / month
platformcustom

Platforms charge for flexibility. Past ~4k leads/month of stable workflow, you're renting a for-loop.

Get this as a model you can edit

A working spreadsheet, not a screenshot: every assumption in its own cell and every result a formula, so swapping in your real quoted rates recalculates the whole thing. Plus the one-page build-vs-buy checklist — core-vs-commodity, unit economics at target volume, rate of change. Downloads immediately.

Stored: what you type above, plus any campaign tag and referring page already in this URL. Not stored: your IP, cookies, or any device fingerprint. Used to email you this one thing and possibly follow up once — not added to a drip sequence.

How this works, and every number it assumes

Two cost curves. The platform charges a plan fee plus credits per row, so its cost rises with volume forever. A custom system carries a fixed monthly cost — infrastructure and the engineering time to keep it alive — and a marginal cost per lead that is close to nothing, because scraping a Google Business Profile is close to free. Fixed-plus-cheap beats zero-plus-expensive, but only after enough volume to absorb the fixed part. Before that point, building is the wrong call.

The fallback slider is the honest variable. Paid data is only bought for records the free source missed, so that percentage drives most of the marginal cost of the custom path.

Assumptions, unedited. Argue with them — that is why they're here.
Platform plan$349/mo
Credits included10,000
Overage per credit$0.020
Custom fixed cost$400/mo
Scrape per lead$0.002
Paid fallback per record$0.03

These are representative, not quoted. Your plan tier, your negotiated rate, and your engineers' time all move the breakeven. The shape of the curve is the argument; the exact crossing point is only as good as the six numbers above.

Live · 3 free/day

GEO visibility checker

Ask a model your buyers' questions cold, and see whether your brand comes up — or who does instead.

Run it

Right now

Owns growth for a US local-services vertical ($15K MRR, scaled from 0). Building agent workflows aggregating spend and lead data across 60–100+ Meta ad accounts into a daily tracker with automatic drift flags.

03

Writing

What I think, not just what I shipped. One argument per post.

2026-08-03 · 5 min

You're renting a for-loop

Agent pricing per row is the most expensive way to do work that never needed judgment. Here's the rule I used to decide what to keep renting and what to own.

Read

Archive

All writing

Notes on GTM engineering — enrichment economics, build-vs-buy, and agent workflows that survived production.

Browse
04

What I want next

Growth engineering / AI GTM roles at product companies.

Open to founding-growth roles at early-stage startups.

Email meBook a call

Right now

Owns growth for a US local-services vertical ($15K MRR, scaled from 0). Building agent workflows aggregating spend and lead data across 60–100+ Meta ad accounts into a daily tracker with automatic drift flags.

Email
subhajitthisside@gmail.comFastest for anything specific.
X
@midfunnelmindWhere I think out loud about GTM systems.
LinkedIn
in/subhajit-goraiFull career history, and where recruiters usually land first.
Book a call
calendarPick a slot; it goes straight on my calendar.

Book a call

Pick a slot that works and it goes straight on my calendar — no back-and-forth.

or book in a new tab

Loads Google Calendar when you click, which means Google sees the request. Nothing loads until then — that is why it is a button and not already sitting here.