My AskAI: $5,000 Pre-Sold Before Building, $40K MRR With Two People, After a VC-Backed Startup Died in Covid

My AskAI, an AI customer-support SaaS, stood at $40K MRR (about $500K ARR) at the time of its July 2025 Indie Hackers interview. It processes more than 75,000 support chats a month while the team remains two people, founders Alex Rainey and Mike Heap. Placed among the AI SaaS cases on Small Start, it sits just below Tony Dinh’s TypingMind at $45K a month and SEObot at $46K MRR. AI startups are usually assumed to need heavy fundraising, but Rainey is someone who already ran that road once and lost the company at the end of it. The second time, he reversed the procedure and sold $5,000 of the product before building it.

The predecessor that died at #2 on Trustpilot, and a bumpy runway

Rainey’s previous company, Pluto, was a VC-backed travel-insurance startup. Customer ratings were strong enough to reach #2 on Trustpilot, yet the pandemic shut down the travel market and the company with it. Product quality and survival are separate questions, and he learned that one in the field.

The restart was not a straight line either. A GPT-3 fine-tuning product shipped in late 2022 topped out around $3,500 in sales over a few months. An earlier tool for writing university personal statements was folded after OpenAI asked them to shut it down. After the loud failure and the small misses, the two sold the fine-tuning product and concentrated on one thing.

Selling before building

What distinguishes the launch is the sequence. By the time they went all-in at the end of January 2023, the pair had collected 50+ registrations for a $99 pre-launch, banking $5,000 before the product existed. The goal-setting matched: “ramen profitability by the end of May,” a floor rather than a moonshot, with proof of willingness to pay as the first milestone instead of a funding round or a viral spike. The stack follows the same philosophy, a no-code Bubble frontend over a Python backend. Ship fast, fix while selling.

Their circumstances at the time were pressing too. Rainey had just let go of his previous product, and Heap was looking for freelance work. With thin means of support, putting demand verification first was rational. The result: $14K MRR in four months. The top acquisition channel at that point was word-of-mouth referral, with organic Google second. Twitter, LinkedIn, Product Hunt, and newsletter ads were all tried, but the quiet channels led from the start. The cost base stayed light as well: OpenAI API costs ran about $600 a month, under 10% of revenue, so the common worry that AI SaaS margins get eaten by compute does not apply here.

Designing away the “why should we switch?” objection

The product’s core decision is to leave a company’s existing support tools in place. My AskAI is built as an AI agent that plugs into Intercom, Zendesk, and similar tools, so customers add AI answers without migrating anything. As a reader in the interview comments put it, the design sidesteps the “why should I switch?” objection entirely. The growth channels mesh with that choice: bottom-of-funnel SEO aimed at readers in the final stage of comparison, content organized around the names of the tools it integrates with, and onboarding calls run as sales opportunities.

It is the same shape as Perfect Wiki weaponizing “you never have to leave Teams”: the fiercer the AI SaaS competition, the more adoption friction, not feature difference, decides the purchase. That two people can carry 75,000 monthly chats is possible because the product absorbs the support work itself. The business effectively runs a live demo of its own product every day.

The uncomfortable side, written down

The numbers have limits of provenance. Both the $40K MRR and the 75,000 chats are the founders’ statements in Indie Hackers interviews, with no third-party verification. And the business stands on top of OpenAI. Light costs are a strength, but within that same dependency the pair have already folded one product at OpenAI’s request. Platform terms, pricing changes, and feature internalization hang over every AI SaaS of this size. Customer-support AI is also a crowded field, and competition with support platforms building AI in as a default is unavoidable.

The revenue ladder of AI SaaS

Order the AI SaaS cases by monthly revenue and My AskAI’s $40K sits below TypingMind’s $45K and SEObot’s $46K, with PhotoAI at $132K a month with zero employees further up. All share one-or-two-person operations, and in this price band headcount and revenue do not move together. The one general rule this case yields: what needs verifying is not whether you can build it but whether anyone will pay, and selling before building is the fastest way to verify it.

Fifty people paid $99 in early 2023

Selling before building worked because buyers already felt the weight of their support queue, and the value fit in one sentence without a demo: add AI answers to the Intercom or Zendesk you already use. Put the same $99 pre-launch in a domain where nobody feels the pain, and what you collect is curiosity, not prepayments.

The ratio of two people to 75,000 chats a month also depends on the product absorbing the support work itself. A business where humans answer the chats never gets that ratio to begin with.

The timing tailwind still has to be subtracted. AI SaaS in 2023 rode strong “we want to do something with ChatGPT” demand, and the sight of 50 people lining up at $99 belongs to that moment. Add two founders seasoned by a failed company and a small product sale, and tracing the same steps does not produce the same curve.