ask what they spent, not what they raised... bootstrapping a b2b software company with ai

i got a guest's intro wrong this week, live, with about thirty people listening.

he co-founded an ai dev company with one other person. two founders, two cities, and that's the whole team. before the taping he'd sent me a prep doc with his corrections and his numbers, the way the best guests do.

on air i introduced him as having built chatbots for some big-name tech brands. that work was real. but it was done at previous employers, for their clients. not at his company.

he'd flagged it in the doc. i read past it. he didn't correct me live, which says a lot about him. so i'm correcting it here. that one's on me.

and the mistake turned out to be the lesson, because the story he wanted to tell was better than the one i told for him.

two people. four hundred and seventy five pull requests per person per month. soc 2 and hipaa both done within the last year. their largest customer brings in somewhere between one and two billion dollars a year.

and they've raised about eighty thousand dollars. none of it institutional. no round, no deck.

a public database listed their funding as zero.

so if you looked them up the lazy way, you'd see a company that doesn't exist. if you asked the right question, you'd see a two-person team selling to companies most startups only put on a logo slide as a dream.

he told me the question he wanted. ask about profitability, not the funding number.

i like that he pushed for it. most founders would let the zero sit there and hope nobody noticed. he did the opposite. he handed me the better question in writing and trusted me to use it.

that is what a prep doc is for. it is the guest telling you, ahead of time, which story is true and which one is just easy to tell.

think about how strange that request is in our world. founders get introduced by what they raised. it's the first line on the slide, the first line in the article, the first question at the dinner. the number becomes the identity.

but the funding number is the laziest question you can ask a founder, and as he showed me, the databases get it wrong anyway. what someone raised tells you how good they are at raising. what they spent, and what came back, tells you how good they are at building.

if you're bootstrapping a b2b software company with ai right now, this is the window. two people with the right tools can ship like a team of twenty. the compliance work that used to need a department got done by a couple of founders in a year.

his thesis went one layer deeper. the harness outlives the agent. models will keep upgrading underneath you. what has to survive every upgrade is the product intent and the customer evidence... what you're building and why, and proof from the people using it. keep that somewhere a model swap can't touch.

it's like a restaurant where the chef changes every season. the menu shifts. the recipes the regulars love, and the reason they keep coming back, stay written down in the kitchen.

long term, i think the most interesting companies of this decade will look like his. tiny, profitable, selling to giants, invisible to the tools that track who raised what.

and the people interviewing them, me included, need to learn to ask better questions.

if you're building something like that and you'd rather be asked about the real numbers, come on the show and send me the doc. i'll read it closer this time.

so when someone asks what you've raised... what number would you rather they asked about instead?