The $1M+ Solo Grok Bot Business (Full Course)

Nick presents Bops, a business-operations agent platform he built for himself after finding existing agent products unsuitable. He describes it as a Grockbot clone and says the build took 36 hours over one weekend, in a single long-running T3 conversation with Opus 5.5 on high mode and without a GitHub connection. A video of Bops received 75,000 views on Twitter, indicating interest in the demo but not, by itself, evidence of customer demand or a million-dollar business.

Bops combines agents with dedicated computers, email, phone numbers, and software connectors. In the demo, a chief-of-staff bot could handle calls and texts, use email, and work on its own computer; each bot had a computer with four screens for multitasking. Nick’s stack included Orgo computers, Agent Mail, Agent Phone, and Composio connectors, with Fly and Vercel also mentioned as hosting options. For building, he recommends researching comparable agents, producing several design variations before combining preferred elements, using screenshots and references, and having agents consult relevant documentation and API references before implementation.

The commercial premise is to configure agents around a specific industry’s workflows rather than sell a generic assistant. A LinkedIn agent, for example, might identify prospects, connect and message them, book appointments, and prepare call briefs. Nick says this requires learning a customer’s bottlenecks and pain points; for small and midsize businesses, he sees a bot with its own phone number for calls and texts as an obvious use case. He suggests that a system could cover tasks such as email, support, billing, calendars, and appointments, potentially using 15 bots.

Nick proposes charging a business $25,000 for implementation, possibly split between the start and completion of a 30-day project, with an optional $3,000 monthly maintenance retainer. He compares the system’s potential value with a front-desk employee costing $50,000–$70,000 a year. Greg challenges the economics from a dental-clinic buyer’s perspective, asking how the $25,000 fee plus $3,000 monthly would produce $51,000 in value; the discussion does not establish a specific return calculation. Nick says selling an owned system rather than requiring a retainer can ease lock-in concerns, and suggests refunding the monthly fee if agreed outcomes are not delivered. He also reports selling a Hermes agent to a business for $5,000 per month, while noting that this price can create friction.

The speakers see two possible business models: custom, high-priced systems for a small number of clients, or a more standardized product sold to many businesses with seat-based pricing and expansion revenue. Greg estimates that industry-specific solutions could command 10–20 times the price of generic ones, and argues that basic software prices may fall while agents and human intelligence become premium offerings. They also note that many buyers may need a demonstration to understand what computer agents can do, and that retention depends on continued value and trust in the provider.

For finding customers, Nick advises testing several industries rather than committing too early to one niche. He recommends contacting 50 people in a personal network who own businesses or know business owners, then building for a receptive prospect. His own conversations led to a first client in promotional products, an industry he had not previously known. He suggests doing the first project free in exchange for a case study, observing the client’s work, and asking them to explain manual processes that might be automated. Greg recommends asking a satisfied client to introduce three local colleagues in the same industry; possible offers include describing the free service as normally worth $50,000 a year or providing a $5,000 credit for introductions. The approach relies on warm referrals, while the proposed financial opportunity remains dependent on delivering clear, sustained value.