You Can Sell a Local Business Service Before You Build the Screens

Payment comes before screen-building

AI can produce landing pages and screen mockups. But a completed screen is not proof that a local business will pay. The first thing to test is not whether you can build it, but whether a business owner who hears the real price will commit time and money.

Local businesses are not customers who can wait indefinitely. According to Statistics Korea (통계청), businesses founded in 2017 had a five-year survival rate of 34.7%, while the accommodation and food service sector had a rate of 24.4%. In other words, only three or four out of ten businesses overall—and two or three out of ten in accommodation and food service—were still operating five years later.

You can distinguish the strength of purchase intent through actions. Sharing information and asking for a quote are stronger signals than saying, “It looks good.” Confirming a launch date and paying a deposit are stronger still. The clearest signal is a decision to keep paying after a paid pilot.

So for one month, what you need is not a finished product. You need a public directory of businesses, a phone, a logbook, and a one-page proposal with a price. If you take payment in advance, state the start date, scope of work, cancellation terms, and refund terms before payment. Korea’s Act on Consumer Protection in Electronic Commerce (전자상거래 등에서의 소비자보호에 관한 법률) generally provides a seven-day withdrawal period for online consumer transactions, but its application can differ when contracting with businesses for business purposes. Review the contract terms separately.

Do not collect contact details carelessly. Use a business’s shared phone number first. Treat an owner’s personal mobile number, or a number combined with their name, as personal information, and manage the purpose of collection and retention period. Remove any business that asks not to be contacted again immediately.

If you sell to one trade for one month first

Consider a solo operator who wants to sell a no-show and repeat-visit management service to one-person hair salons in Bupyeong District, Incheon. The proposal is: “I will run pre-appointment reminders and post-visit return reminders for KRW 99,000 per month.” This is not a market-rate claim. It is an experimental price for testing purchase response.

The operator moves between maps and booking pages, recording each salon’s name, shared phone number, booking method, and closing days. After each call, they write up the conversation and separately track promised messages and follow-up dates. If they build screens first, they can easily skip this process and spend time on features they imagined.

In the first week, narrow the focus to a place and trade, such as 30 one-person hair salons in Bupyeong District. On calls, ask how many no-shows they had in the past month, how they currently send reminders, and who spends how much time on the work. Do not ask, “Would you use this feature?”

In the second week, use AI to summarize call notes and group recurring phrases. Do not hand over the calls or judgment itself. If an owner says, “Filling empty slots with returning customers is more urgent than no-shows,” do not cling to the original problem. Change the proposal.

In the third week, offer a paid pilot with a real price and start date. For four weeks, manually provide reminder copywriting, recipient-list preparation, and repeat-visit tracking, while collecting a deposit or the first month’s fee. Put the scope of service, information the customer must provide, and cancellation and refund terms on the application page.

In the fourth week, deliver before you automate. When a person directly sends messages, organizes recipients, and reports results, you learn where information arrives late and which wording blocks approval. At the end date, do not only ask about satisfaction. Ask whether the business will continue next month at the same price.

Much of the work still belongs to people. Calling outside busy hours, understanding an owner’s language, and earning enough trust to handle a customer list cannot be replaced by a screen. What competitors will struggle to copy is not the feature. It is the accumulated data on which trade rejects which price, direct customer relationships, and the trust and distribution routes that can lead to referrals.

Cases that called first and built later

French restaurant service Zenchef

Zenchef’s predecessor, 1001Menus, began by helping independent restaurants with websites and reservation management. Its two founders called restaurants from the phone directory from beginning to end and showed non-working screen mockups on an iPad. In their first summer, before they had a product, they signed 50 paying restaurants.

The number of restaurants then grew from 50 to 350 in six months, and reached 650 after the sales team expanded. The company now says it supports more than 36,000 restaurants across more than 20 countries; this is a company-reported figure. What it secured early was not only reservation technology, but also a route for calling on and visiting restaurants to win contracts.

US food-supply ordering service FareFood

Before building software, FareFood’s founder personally called 47 restaurants in the Tampa Bay area of the United States. The initial menu-analysis idea received little response. Instead, the founder repeatedly heard that comparing prices across food suppliers and combining orders was the bigger problem. The team learned that its first idea was wrong before building screens.

FareFood became a service centered on that problem, offering price comparison, consolidated ordering, invoice management, and payment management. The company’s reported savings figures need separate verification, but the 47 calls clearly changed what it built. Over time, accumulated food-supply prices and order data become harder to copy than screens.

US customer onboarding management service GuideCX

GuideCX’s founder first contacted 76 managers responsible for implementing enterprise software. The founder asked the same 16 questions, showed screen mockups, and proposed pilots intended to convert into paid use. Of the 76 prospects, 14 committed to participate, and 9 paid after the free pilot ended.

According to a founder interview, annualized recurring revenue reached about KRW 560 million by the end of 2018. This figure is based on the company’s account, but the sequence of measuring contact volume, pilot commitments, and actual payments is worth studying. The trust and work records gathered through interviews became the starting point for the later product.

Things you could build from this

1. A trade-specific call validation log

This is a service for someone selling no-show management to hair salons or order organization to neighborhood side-dish shops. It keeps a business’s shared phone number, current workaround, proposed price, reason for rejection, and follow-up date in one place.

AI can summarize call notes and group recurring objections, making it manageable for a solo operator. The first screen should show local business lists, statuses such as “not contacted,” “call completed,” “price proposed,” and “paid start,” plus the businesses to call again today.

2. A paid-pilot application page builder

This is a service for a two-person business testing a new offering, such as custom signage or store-promotion management, to create an application page with a price and clear terms. It goes beyond a simple consultation request and leads to deposit payment for a service with a stated start date and scope.

Screens have become easier to build, but missing transaction terms can create disputes. The first screen should include fields for the target customer, work to be done over four weeks, total price, deposit, what the customer must prepare, and cancellation and refund terms, in that order.

3. A manual operations desk for the first ten customers

This is a service for someone running repeat-order reminders for neighborhood side-dish shops that have just started delivery. It records each shop’s customer information, this week’s message recipients, copy that needs approval, order outcomes, and time spent on work.

AI can draft messages and weekly result reports, but people handle approval to send and customer responses. The first screen should show today’s tasks, shops that have not sent their information, and verifiable order and inquiry outcomes rather than revenue by shop.

4. A small pilot marketplace introduced by a local association

This is a space where a local hairdressers’ association or merchants’ association recruits five member businesses at a time for reviewed services. Builders who find it hard to approach local businesses gain an introduction route, while associations can offer members clearly stated options instead of random sales calls.

In markets where trust and distribution matter more than technology, relationships with associations become harder to copy as they accumulate. The first screen should show the association name, eligible trade, service scope, price, remaining slots, and reviews from businesses that actually used the service.

Why this matters where you are

You can test this approach in your own market by checking whether local businesses will discuss a concrete problem, a start date, required information, or payment—not merely praise an idea. Consumer rules, business-contract terms, and how contact information may be used will differ by market, so verify those before taking prepayment. The useful action is the same: sell and deliver a narrow paid pilot before deciding what to automate.

What to check today

Call five publicly listed business phone numbers over 30 minutes and state the same price. If you have a concrete conversation about the problem with at least two businesses, and one of them asks about a start date, required information, or payment method, it may be worth preparing the next proposal. If you receive only praise and no commitment to an action, do not build the screen. Change the problem or the trade first.