Loading page…
Before shipping an AI selling agent, they ran a staffed e-commerce sales floor in Arizona where people conversationally sold for brands by hand — millions of real conversations that became the training set no competitor can buy
7.5 years old with all three cofounders still full-time · previously ran SMS Sales, a staffed "e-commerce sales centre" in Phoenix, Arizona, producing "millions and millions" of manual selling conversations · three AI products, all under a year old · single-digit-percentage of customer website traffic now arriving from consumer AI chat destinations · statutory exposure of $1,500 per non-compliant SMS under the US Telephone Consumer Protection Act
"We had millions and millions of manual sales conversations previously. It's been an unbelievable set of training data." Three and a half years before the AI selling agent existed, they staffed an office in Phoenix with humans doing the job by hand — that archive is the moat, not the model.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
Postscript is an SMS marketing and conversational-AI platform for e-commerce, roughly seven and a half years old at the time of the interview, with all three cofounders — Alex Beller, Adam and Colin — still in the business full-time. The interviewer's figure of about 20,000 brands is confirmed in passing; no revenue is ever disclosed. Beller's framing of the journey is the useful part. For years it was, in his words, "a scaling SaaS 2.0 business with playbooks that have been established for a long time." Then the LLMs arrived and, as he puts it, hit reset — "this is a new world, everything's being reset, and so we're going to have to reinvent the company all over again." Three AI products came out of that, all less than a year old: Brand Center, Infinity Testing and Shopper. He is candid that the market was never easy: they have competed the whole way against better-funded, bigger competitors, and in the early years he would get on sales calls where most prospects said they were never doing text messaging. He is equally candid about the answer to whether he would start now. Asked directly, he says it is too late within his space — and that he would absolutely do the journey again.
The earliest acquisition detail Alex Beller gives is a sentence about rejection: when they started, he would get on sales calls and most prospects said they were never doing text messaging and that it was not going to work.
The quick-reference card — acquisition channels, replication playbook, and risk map — is members-only.
Join the Unicorn Club — $5/moCancel anytime, in one clickPaste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself.
Data credibility: Friendly CEO interview on a technically-oriented AI show; no adversarial questioning and no data shown on screen. ⚠️ No revenue figure of any kind is disclosed — no ARR, MRR, growth rate, churn, retention or funding history. The customer count of about 20,000 brands was supplied by the interviewer mid-sentence and confirmed by the CEO in passing rather than volunteered, so treat it as approximate. Founder-stated and unaudited: 7.5 years old, three cofounders all still full-time, three AI products all less than a year old, a prior business called SMS Sales run from a staffed "e-commerce sales centre" in Phoenix, Arizona, "millions and millions" of manual sales conversations from that operation, single-digit percentages of customer website traffic now arriving from consumer AI chat destinations, and $1,500 per non-compliant message under the US Telephone Consumer Protection Act. The named customer brands (Hexclad, Brooklinen, True Classic, Dr. Squatch) are given as illustrations. The Dr. Squatch "Chad mentions his dog" result is described qualitatively with no lift figure attached. Beller repeatedly notes he did not personally architect the systems he is describing — on the RAG implementation and the data-reconciliation layer he says explicitly that he does not know exactly how it was implemented under the hood — so treat the technical description as a CEO's account rather than an engineering one. Cofounders are identified only by first name (Adam and Colin), as is the legal team (Elliot and David).
No revenue figure of any kind appears in this interview — no ARR, no pricing, no growth rate — so the first dollar has to be read from how Beller describes the value.
The flywheel that matters here is data, and Postscript built it by hand.
Postscript's version of this was an adoption curve rather than a silent launch. For years Beller was getting on sales calls where most prospects said they were never doing text messaging, and the company spent seven and a half years competing against better-funded rivals while that belief slowly changed.
The rest of this breakdown — 9 more sections — is in the Unicorn Club, with every section of all 1,335 ideas: first customers, first dollar, the flywheel, and the silent stretch.
Join the Unicorn Club — $5/moCancel anytime, in one click