Adam Young

co-founder of impact driven, building cruxops(.ai)

[Adam Young @whyg0ng]

@mubashirtasinn Thanks for the comment Tasin! For me, being earnest was the part that surprised me most and I personally think it’s the hardest part for many founders

[Adam Young @whyg0ng]

6 lessons I've learned during early sales:

0/ learning should come first. revenue is the byproduct.

early founder-sales is not primarily about closing as much revenue as possible.

it’s about learning:
- who actually cares
- what problem is real
- what makes them respond
- what makes them buy

revenue comes as a byproduct of getting those answers right.

1/ narrowing down the target DOES matter.

I used to target any 30–200 person company selling B2B software or AI.

turns out, you learn almost nothing by sending the same sequence across all of them.

separate targets by their characteristics + signals.

each segment should feel stupidly small.

2/ compounding matters.

if engineering has CI/CD, early GTM should have CS/CD (Continuous Sales & Continuous Distribution).

whether you’re building, sleeping, or busy doing something else, keep thinking:

"how can the work I’m doing today keep producing results tomorrow?"

3/ find the fastest way to learn. not the way that looks right.

"let’s send cold emails for a few weeks, run some ads, then analyze the results."

naaaaah.

want to know what leaders think about your vision? ask them so.

want to know what ICs think about what you’re building? call them and ask.

early stage is about shortening the feedback loop as much as possible. (and there's never a "right" answer for that)

4/ performance marketing is mostly a waste of money in early sales.

paid acquisition makes much more sense once you know who wants the product and why.

before that, you’re basically paying to amplify something you still don’t understand.

communities, SEO/AEO/GEO, founder contents, cold outreach... whatever.

find the cheapest channel that helps you learn + distribute at the same time.

5/ be honest & earnest. show your passion.

you don’t need to make yourself look bigger than you are.
you don’t need fake enterprise language.

- say what you’re building.
- say why you believe it should exist.
- say why you care!

especially early on, people often buy into the founder before they fully buy into the product.

[Adam Young @whyg0ng]

@vict3rr Same thing he did on preseason

[Adam Young @whyg0ng]

this is what happens when you trust AI to write your emails

[Adam Young @whyg0ng]

getting the hard stuffs done.

[Adam Young @whyg0ng]

try it out now, no credit card needed!

https://t.co/uEJUjNZ0hK

[Adam Young @whyg0ng]

introducing Claude for prospect research.

drop in a prospect URL and get the research that matters in minutes.

no AI noise. no 10+ tabs. no digging for the angle worth leading with.

[Adam Young @whyg0ng]

@F1 glad to see him smile again on P1!

[Adam Young @whyg0ng]

@SpursOfficial why won't they just stay and train more

[Adam Young @whyg0ng]

@liftplaycode Average cold caller life lol

[Adam Young @whyg0ng]

@vercel get your Prospect Research done the way you would

https://t.co/o2o9gFnt9H

[Adam Young @whyg0ng]

@vercel went from 10 sdrs to 1, with AI replacing the work of the other 9. how id they actually pull this off?

here are 5 lessons you should note:

1/ building on a weak foundation will end in failure
- having something manually done + accumulated as first-hand experience matters.
- if your GTM motion is random and you cannot make it repeatable, building AI on top of it will almost always go to waste.

2/ start small
- even if you already have a repeatable motion.
- Vercel started by automating its inbound sdr work first. (not outbound BDR work, not live demos...)

3/ building is not the hard part. the discipline is
- a GTM engineer spent part of 6 weeks shadowing one of Vercel’s best-performing inbound SDRs.
- he documented every decision point, data lookup, and reply format the SDR used to qualify leads.
- And instead of asking AI to simply "qualify leads like our best rep", he turned those individual steps into a deterministic workflow.

4/ iteration matters (ft. human in the loop)
- once the framework for a single workflow was figured out, they applied the same approach across roughly 30 different workflows.
- hence, trusting AI 100% was never the answer.
- every new workflow needs someone for reviewing the output + feeding corrections back into the system.

5/ build for production! (ft. tokens aren't free!)
- Vercel spent around $5k/year across infrastructure + tokens (tiny compared to the scale they operate).
- this only works if the architecture was designed for scale.
- if a single prospect research costs you $2, the system would brutally fail.

[Adam Young @whyg0ng]

@HotspurLane 2/10 whenever he took his shirts off after scoring, we dropped crucial points for us

[Adam Young @whyg0ng]

@danielcberk these people just sending dms with whatever hooks ai suggests. no effort there.

[Adam Young @whyg0ng]

most founders define their target after fully building the product. then find out no one wants it.

they pivot, then call themselves lean.

[Adam Young @whyg0ng]

@devWalo Just followed you! Let’s gooooo

[Adam Young @whyg0ng]

founders, you finally launched your product.

how are you getting your first 10 users?

[Adam Young @whyg0ng]

@CrocDJL @HotspurLane yes but richy gets 90k a week and we've got a lot of money..

[Adam Young @whyg0ng]

receiving cold outreach and seeing phrases like:
"resonated with me...", "curious: ...", "any interest in...?"

makes me not want to read the rest. (and I definitely won’t reply)

[Adam Young @whyg0ng]

@papematarsarr15 @juventusfc Pape, sorry what spurs have done for u in the last moments. But as a fan, I just wish u all the best for the future