Loading page…
A bootstrapped software founder used a profitable OCR product and customer knowledge to start a broader AI company, with about $250K annual revenue reported across the existing business.
Tag Team was early and had no established revenue figure
Use the first focused product to learn who pays, then expand only when customer requests point to a related problem.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
CK Lee built Taggun, an OCR software business, as a part-time venture while holding a full-time engineering job. The host says Taggun grew over five years to about 50 customers and roughly $250,000 a year in US-dollar revenue, close to half a million New Zealand dollars. CK described monthly revenue in the $17,000–$20,000 range at one point and later cited about $400,000 in revenue over the last 12 months, so the interview’s run-rate figures are not perfectly consistent. Taggun had a mix of smaller customers and enterprise accounts, with roughly half of revenue attributed to enterprise and several customers requiring dedicated hosting. CK then started Tag Team, an AI-focused venture intended to build multiple products around related machine-learning problems. The new company was early: it had launched on Product Hunt and was still seeking product-market fit. The conversation covered a small Google Ads budget for Taggun, organic search, customer interviews, and a possible seed round to hire more engineers and sales capacity.
For a technical B2B product like Taggun, the first customers are usually found by identifying a narrow operational problem and speaking directly with the people who own it. CK’s history points to an existing engineering background and a focused OCR use case, then gradual customer accumulation over five years.
The quick-reference card — acquisition channels, replication playbook, risks & traps, and our verdict — 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: Nathan Latka interviews CK Lee and reports Taggun at about 50 customers and approximately $250,000 annual USD revenue. CK later cites $17,000–$20,000 monthly revenue and about $400,000 in revenue over the most recent 12 months; the transcript does not reconcile these claims. Customer lifetime value, ad conversion counts, and enterprise share are also spoken estimates, with some garbling. No third-party financial records or dashboard are shown. Tag Team is described as early-stage, with no established revenue number in this conversation. The profile separates Taggun’s traction from Tag Team’s status.
The interview focuses on Taggun’s later revenue rather than describing the exact first sale, so an early founder should use a direct paid pilot to make the first dollar legible.
Taggun’s reported growth combined recurring B2B usage with search and paid acquisition experiments.
A developer can spend months improving an API while the market stays quiet, especially when the buyer is difficult to identify.
The established business, Taggun, provides OCR capabilities for software customers who need document recognition. Its customers range from low-price monthly users to enterprise accounts, and some enterprise buyers requested dedicated hosting for security or internal policy reasons. Tag Team is CK’s newer venture, intended to build AI products around adjacent machine-learning needs. The conversation presents it as an early platform direction rather than a mature unified suite. One customer request led to a prototype for a classification problem, while the founders were exploring whether users of one product would adopt others. That is a meaningful product question: shared customer access does not guarantee shared jobs-to-be-done. A credible expansion strategy starts with a specific adjacent workflow and validates that the same buyer has budget and urgency. The technology is changing quickly, and CK noted the availability of more public models and datasets than five years earlier. That may lower prototyping effort, but enterprise reliability, hosting, security review, and integration work remain substantial product requirements.
Taggun monetizes through B2B software usage, with customer pricing ranging from about $1 per month to enterprise contracts. The interview reports around 50 customers and roughly $250,000 in annual US-dollar revenue, while later discussion cites monthly revenue of $17,000–$20,000 and roughly $400,000 over the latest 12 months; those figures do not reconcile cleanly and should be treated as approximate creator-reported data. CK says about half of revenue came from enterprise customers, with three to six large accounts mentioned at different points. Dedicated hosting was one requirement for customers with internal security reviews. The acquisition discussion includes an approximate $2,000 customer lifetime value and a $50 daily ad budget, but the transcript is garbled around conversion definitions. Tag Team itself had not established comparable revenue in the conversation. A new founder should keep the two businesses’ sales and costs separate, identify usage and hosting margins by account, and avoid applying Taggun economics to the newer product portfolio until it has its own paying customers.
The company’s technical stack and exact hosting costs are not specified in the interview, but the commercial requirements are visible: OCR and machine-learning software, integrations for B2B customers, Google Ads, organic search, and dedicated hosting for some enterprise accounts. CK described the current AI ecosystem as having more accessible public models and datasets than when Taggun began five years earlier. That can shorten experimentation, but it does not eliminate the engineering burden of reliability, data security, model evaluation, or support. The business has run with a small combined team, described as three people including CK and Hannah, and CK was considering hiring three more engineers and one senior sales person if funding arrived. Model each customer segment separately: usage-based compute for smaller accounts, and hosting, implementation, and support costs for enterprise. The reported customer lifetime value should not be used as a payback calculation until churn period, gross margin, and acquisition attribution are clear. Keep ad tests capped and stop keywords that attract end users instead of software buyers.
Copy the focused bootstrap discipline behind Taggun: build around a clear technical use case, earn customers over time, and use paid search as a measured experiment rather than an open-ended growth lever. CK’s advice to keep learning from users is especially useful when moving into a new category. Change the temptation to bundle several AI products before proving that customers want a shared platform. Tag Team’s early Product Hunt attention and one customer prototype are starting signals, not evidence that cross-sell will work. Pick one customer type and a single high-value workflow, then sell the first implementation before expanding the roadmap. Use keyword tools to develop hypotheses, but qualify each search term by buyer intent and measure paid usage rather than clicks. If pursuing venture funding, articulate why the broader opportunity supports venture-scale growth and how the initial product creates a repeatable distribution advantage. A profitable niche product and a venture-backed portfolio can have different operating goals; be explicit about which one the team is building.
The main strategic risk is confusing a promising adjacent product idea with evidence that a multi-product company will compound. Tag Team was early in the interview, and CK acknowledged uncertainty about whether customers of one product would use another. Enterprise OCR also carries concentration and delivery exposure: a few large accounts can represent a significant share of revenue, while dedicated hosting adds operational work. The revenue figures in the conversation are inconsistent, so the quarter-million annual figure should be treated as approximate rather than as a reconciled financial statement. Paid acquisition may also plateau if the addressable search volume is limited; the host raised that concern when asking why the team did not increase its daily budget. Finally, AI models and libraries are improving quickly, which lowers barriers for competitors and can make yesterday’s technical edge less durable. Defend the business with integration quality, customer-specific reliability, clear security practices, and a focused buyer problem. Do not hire ahead of validated demand simply because a funding round is possible.
The four founder playbooks — first customers, first dollar, the flywheel, the silent stretch — are in the Unicorn Club, on this idea and on all 1,403 others. The story, the numbers, the sources and every later section stay open to everyone.
Join the Unicorn Club — $5/moCancel anytime, in one click