Ivan Nedelkovski Built Lancer from the Upwork Sales Work He Kept Repeating
How a tool that began with the repetitive Upwork sales work he did while running a development agency opened early sales through a coach's referral, charged by sales conversations started, and grew.
A Tool Born from Repeated Sales Work
Ivan Nedelkovski grew MVP Masters, the software development company he started in 2020, to 15 employees. In 2025 he sold the company and moved into the product business. Lancer started from the customer acquisition work he repeated while running that company.
In 2024, on a friend's recommendation, he began using Upwork. A 10,000-dollar development project he won there led to a long-term contract worth 15,000 dollars a month.
The problem was the time it took to keep selling. Checking new project postings, choosing the work that fit the company, and writing proposals to send to clients left little room to focus on development. While he was doing other work he missed suitable postings, and even when he handed the task to support staff, the job of making them understand his field well enough remained. Ivan decided to build software that would keep doing this repetitive work.
He also subscribed to a competing service priced at 500 dollars a month. He looked at whether existing customers were paying for the problem, whether there was room to improve the product, and how to approach customers.
He built the early version over a weekend, and acquaintances who used it won contracts too. Two people then worked full-time for six months to complete the first commercial version.
Automation for Picking Postings and Writing Proposals
Lancer's job begins with finding project postings. It narrows candidates by budget, the client's past spending, the actual hire rate, and feedback records, and then AI reads the posting's content and judges whether it fits the user. Even postings that carry the same technology name differ in the work they require and the client's conditions, so judgment about the content is needed even after a posting passes the search conditions. Narrowing the sales target at this stage is what makes the later proposal writing meaningful.
For proposals, the product draws on the specialties, work experience, and service description the user has registered. Instead of sending the same introduction to every posting, it composes content connected to that project and answers the questions the client raised. Users can review or edit the text before sending, and cancel applications that do not fit. The product widened its automation while keeping a point for the user to step in.
At first the company linked Upwork accounts obtained separately to the customer's agency. It could not see incoming messages, and multiple customers were tied to the same account. After four months and several thousand dollars, it switched to connecting the customer's own account directly.
Account Access and Platform Policy
In a business like this, the approach to account access is tied to the product's survival. Upwork states that it may take action against the use of unauthorized automation tools, including warnings, temporary account restrictions, and permanent bans. A product like Lancer therefore has to solve not only whether the automation works technically but also whether the way it works is permitted on the platform. The fact that the customer's sales base sits on an external platform does not change as the product grows.
Early Sales Opened by a Coach's Referral
Early sales were unlocked by an affiliate referral from an Upwork coach. According to Ivan, the product reached 10,000 dollars in monthly recurring revenue (MRR) with about 30 paid customers, and many of them came through this coach's introduction. The coach was already someone who dealt with customers in the same field. The product reached customers who were likely to buy, through a person they trusted.
Charging by Sales Conversations Started
The billing basis was set not on how many proposals were written but on how many conversations with clients had started. On the price list published in September 2026, under quarterly billing, the basic plan works out to 99 dollars a month and includes 15 opportunities with a conversation started each quarter, with 19 dollars charged for each one beyond that. The higher plan works out to 333 dollars a month, removes the cap on that count, and lets the customer connect multiple Upwork accounts. By making a client's reply to a proposal the billing unit, the design lets customers judge cost against sales results.
Helping customers became a product as well. When the team helped with setup directly, results improved, but supporting every customer became too much work. Including this support in higher plans also raised revenue per customer. The public price list bundles setup assistance, strategy consultations, and periodic check-ins, which shows the shape of a software license combined with operations support.
Selling Lancer with Lancer and Meeting Customers Through a Free Tool
Ivan also used Lancer to sell Lancer. He pitched the product to clients who were looking to hire someone to send proposals on Upwork, and the 45-day results he published were 149 prospects contacted, 23 consultations, 17 new subscribers, and a 5,083-dollar increase in monthly recurring revenue. In the consultations he showed that Lancer was the one that had contacted them first. The process that produced the consultation became a demonstration of what the product does.
A free tool, UpworkMRR, also served as a channel for meeting customers. Ivan built a service that analyzed more than 300,000 public profiles to compare freelancers' income and contract records, and the site carries a link through to Lancer. Freelancers curious about their own results and standing in the market overlap with Lancer's potential customers. He also used the same data to select which freelancers to contact for sales.
As subscription revenue accumulated, acquisition offers for the whole business arrived too. Ivan said that after connecting his payment account to TrustMRR, he received a 150,000-dollar acquisition offer within hours. As recurring revenue attached to software two people had built, the options of continuing to run the product or selling the business appeared together.
Dividing Roles and Extending Through External Integration
In TrustMRR's payment-linked data for September 24, 2026, Lancer showed 21,498 dollars in monthly recurring revenue and 124 active subscriptions. It is a case of building monthly recurring revenue in the 20,000-dollar range from a relatively small number of subscriptions, before winning a large base of signups. It can be read as the result of combining a product fitted to the customer's sales work with pricing in the hundreds of dollars a month.
As operations continued, the founders' roles also settled. Ivan said he would spend more time on growth and sales, and that co-founder Tome, who had handled technical development from the start, would take charge of the product overall. After focusing on improving the product, there was a need for someone dedicated to telling customers about those improvements and turning them into purchases. The two-person team also moved to a stage where development and sales each had an owner.
The product also expanded toward letting customers connect their own automation. In Lancer's external integration mode, Lancer handles finding postings and the final submission, while writing the proposal or deciding whether to apply can be done by the customer's program. Because the structure passes posting information to an external system and receives the result back, customers can insert their own criteria or a separate AI in between. Beyond a tool used in a fixed screen, it became a product that can sit inside another sales system.
The part worth noting in Lancer's business structure is that the scope of automation and the billing unit are connected. The product takes on the repeated work of reading postings, choosing targets, and sending proposals, and the price is tied to conversations with clients, a stage of the sales process. Through this, customers can judge the fee by the time they save and the opportunities they gain. The opportunity to build a small business with generative AI lies in work customers already do and whose economic value can be explained once it improves.