The $2.7K/Day AI Business That Practically Runs Itself

Business summary

Aaron Edwards, founder of DocsBot (rendered as “Docspot” in the subtitles), built an AI customer-support agent business after testing eight ideas in six months while keeping his job. The company helps businesses automate support questions using their documentation and knowledge bases. Its business model has expanded from self-serve software sales to working with agencies and implementers who set up and maintain AI agents for small and midsize businesses.

The opportunity Aaron emphasizes is not simply building AI software: many businesses lack the time or expertise to adopt it. A service provider who identifies a business problem, implements an agent, and demonstrates its value can charge substantially more than a software subscription alone.

Business playbooks and actionable tactics

• Find a problem through firsthand exposure. Aaron saw his former employer’s support team answering repetitive questions and recognized that AI could handle a significant share of them. WPMU DEV became the product’s first customer.
• Validate before building. He created a waitlist and surveyed roughly 200–300 business owners about the features they needed, then used the responses to define an MVP that he built over about two weeks.
• Build in public. Posting product demos helped the early product reach a Japanese tech influencer, leading to a viral post and a surge of users. Aaron’s takeaway: a small audience can still produce an important connection.
• Use a paid commitment to qualify users. Rather than offering a free trial, the product asked users to pay and offered a refund if it did not work for them. Aaron said this helped screen out less serious users and reduced support demands.
• Sell implementation, not just software. Agencies can handle setup, onboarding, maintenance, and reporting for businesses that would not adopt a self-serve product on their own. DocsBot supplies the underlying platform, with agency partners acting as sales and implementation channels.
• Price around customer value. Estimate the labor or expense saved, make conservative assumptions, and explain the calculation to the prospect. The interview suggests aiming to deliver about 10 times the fee in value where feasible.
• Automate prospecting and delivery. Aaron described using AI tools and connected data services to find suitable businesses, identify contacts, analyze their support content, build a tailored demo, and generate recurring client reports.
• Prioritize differentiated outreach. Aaron’s suggested lead channels, in order, were in-person networking, phone calls, cold texts, and email. He also recommended targeting businesses hiring support agents or using older support/chat tools, since these can signal existing demand and budget.
• Make marketing memorable. At a conference, the team used AI-designed shoulder-mounted robot demos rather than paying for a booth. The robots served as conversation starters and live product demonstrations.

Product and customer fit

• The core product answers customer questions from a company’s documentation and can escalate unanswered queries to a ticket or live support.
• The platform can also support voice agents, sales lead collection, and internal knowledge tools, such as HR bots. Aaron said the company has focused primarily on customer support, especially cases involving detailed product specifications, policies, or complex documentation.
• Good-fit customers have an existing support team, meaningful ticket volume, and a need for a maintained dashboard and ongoing improvement. Some solo developers and small users cancel because they can build a simpler tool themselves; Aaron described these users as a weaker fit for the business.

Metrics and commercial details

• Company age and growth: About three and a half years in operation. Aaron said the company reached more than $1 million in annual recurring revenue (ARR) and had grown roughly 3–4% per month.
• Profitability and team: Aaron reported gross margins above 90%, no office, and a team of three, including himself, a longtime friend, and his son. He said he typically works around 20–40 hours a week, with substantial time off.
• Software pricing: Plans discussed included $50/month for smaller users, a common plan around $150/month, a $500/month self-serve business tier, and enterprise pricing as needed. Aaron cited average revenue per user of roughly $130/month.
• SaaS retention: He reported revenue churn around 3% and customer lifetime value slightly above $2,000. The share of cancellations attributed to customers choosing to build their own product rose from about 1% to 13% over the prior year, though he said these were often smaller customers.
• Agency pricing: A typical offer is about $2,000 upfront for setup and roughly $500/month for maintenance and reporting. Agencies may pay DocsBot approximately $20–$50 per agent, depending on the arrangement.
• Customer value example: The discussion estimates that support tickets take about five minutes of employee time each. For a business handling 2,000 tickets monthly, automating half could save substantial labor; the speakers use an illustrative estimate of about $7,000 per month in savings versus a $500 monthly fee. Actual savings depend on ticket volume, resolution rate, and labor costs.
• Agent capacity: The platform was described as handling up to about 15,000 tickets per month without additional charges under the referenced plan.
• Early traction: According to Aaron, the business reached about $10,000 in monthly recurring revenue within three months. Earlier experiments included AI profile pictures, which generated about $10,000 over two months before demand faded.
• Conference marketing test: Three robots cost about $50 each to make. At one conference, they generated an estimated 10–20 leads and had produced at least two customers by the time of the interview, with further conversations ongoing.

Revenue clarification: The host’s introduction describes the business as making more than $1 million in annual profit, but Aaron’s specific figure in the interview is more than $1 million in ARR. ARR is not the same as profit; the interview separately reports gross margins above 90%.

Presenters and sources

• Presenters: Chris Koerner (host) and Aaron Edwards (guest), founder of DocsBot.
• Companies and tools discussed: DocsBot, WPMU DEV, BuiltWith, Apollo, TREGT (spelling as heard in the subtitles), ChatGPT, Claude, and OpenAI.