A Tooling Layer Is Opening Between Public Challenges and Follow-on Delivery

Agencies are putting problems on the table first

The Ministry of SMEs and Startups (중소벤처기업부), the Korean ministry responsible for small businesses and startups, is recruiting startups to solve public-agency problems from September 15 to October 6, 2026. It will select about 20 companies from 22 challenges submitted by 18 agencies. The new programme is recruiting companies to carry out those 22 challenges proposed by the 18 agencies.

Selected companies can run field trials with public agencies and receive up to KRW 100 million per challenge to build a prototype, meaning an initial product that actually works. They may also be connected to research-and-development programmes if they meet separate requirements. The maximum amount is not guaranteed for every company, however, and the availability and terms of follow-on support may vary by programme requirements.

The problems are already specific. The National Data Administration (국가데이터처) is seeking a tool to anticipate population density in major areas during festivals and holiday periods. Korea Expressway Corporation (한국도로공사) wants a tool that detects rain and snow in road video so it can respond earlier to black-ice risk. Korea Airports Corporation (한국공항공사) has proposed a baggage-delivery service using real-time airport information.

There are two routes in. One starts with an agency publishing the problems and data it can disclose. The other lets startups propose the use of public data that has not yet been published. Sixty-three companies previously applied through the second route, and the Corporate Public Data Problem-Solving Support Center at the National Information Society Agency (한국지능정보사회진흥원) helps them secure data.

The call guarantees field trials and initial product development, not a formal purchasing contract. Access may be restricted for non-public information involving personal data or security, and for third-party-rights material without valid permission to use it. Agencies also are not always required to process data into the form a company wants. The opportunity, then, is less about the technology alone than about connecting the broken work between challenge discovery, data negotiation, trial records, and formal procurement.

The opportunity appears when the tools change

Consider a hypothetical chief executive of a four-person road-safety video-analysis company. The company has the technology, but does not know which agencies hold which problems and video data. Every week, the chief executive opens government notices, agency websites, and attachments separately to find suitable challenges.

For each challenge under review, the company copies six items from the notice into a spreadsheet: deadline, eligibility, required submissions, contact person, needed data, and success criteria. Once it decides to apply, it enters the same information again in a proposal document. When the notice and its attachments use different wording, it calls or emails the contact person and records the answer separately.

If the video data is not public, the work breaks again. The company searches the Public Data Portal (공공데이터포털) for similar data, writes a data-provision request, receives the agency response, and then reflects it in the business proposal. If it cannot receive the original video, it must negotiate in a separate document whether it can use video with faces obscured, aggregated values by time period, or an approach that analyses data only within the agency.

After selection, meeting records are scattered across email and chat rooms, trial results sit in developers’ log files, and agency confirmations appear in reports. Someone manually collects detection times and errors before and after the trial for the final report. When formal purchasing is discussed, the company must again organise operating costs, the permitted scope of data use, and whether it can sell the product to other agencies. The same facts move through three or four documents from discovery to procurement review.

This programme changes the beginning of that flow. Because agencies put forward the problems they need solved and the data that may be used, startups can begin with disclosed operational problems rather than vague sales outreach. There is also now a route to propose not simply that they want non-public data, but the problem to solve, needed fields, alternative data, and protection methods together.

A small connecting tool can fit here. It could read a notice and bring the six items into one screen, pass the same information into a draft data-provision request when needed, and keep trial goals and results connected after selection. It could then turn those records into an operating plan and materials for a formal procurement review, reducing repeated entry.

Some work clearly remains with people. Agency staff must decide whether they can actually release the data. Frontline staff must test whether the tool fits their work. Privacy and security teams must approve the scope of access. Even a successful trial must pass budget and purchasing procedures. A connecting tool should therefore not replace decisions; it should prevent the records needed for those decisions from being lost.

Other programmes have designed for the path from trial to purchase

CivTech Scotland (CivTech), run by the Scottish Government, has asked public agencies to publish problems rather than product specifications since 2016. Companies first receive funding to explore feasibility for three weeks. If selected for the next stage, they build an initial working product over 15 weeks. The programme does not stop there: products with demonstrated results can move into additional field trials and purchasing contracts.

In a seed-use challenge from a forestry agency, Cumbria Tree Growers participated and continued as Tape4Trees. It automated sowing and transplanting by placing seeds on biodegradable tape. According to the operating agency, it saved about KRW 1.8 billion at the exchange rate used at the time within 18 months and increased planting productivity more than twentyfold. Because field costs and workload were recorded from the start, trial results could lead to larger facility investment.

Innovative Solutions Canada (Innovative Solutions Canada), a Canadian federal programme, lets federal agencies publish problems for solutions that are still difficult to buy in the market. Its Challenge Stream runs proof-of-concept and prototype development through phased research-and-development contracts, while its Testing Stream separately supports trials in real government environments. A government agency that passes a trial can become the first buyer, but purchasing is not automatic: it proceeds through a separate commercialisation contract.

Canadian company Spexi Geospatial began with a challenge from Natural Resources Canada and Transport Canada to collect risk data near railways. It tested an approach using remotely piloted aircraft to collect changes in terrain and facilities along rail sections for about two years from 2021, then received its first commercialisation-pathway contract in 2023. It is a case where a problem call, trial records, and a purchasing decision formed one flow.

Four small things you can build now

1. Public-challenge matcher

• What: Extract deadlines, eligibility, needed data, and success criteria from notices and attachments across agencies, then make them comparable in one place.
• Who: Safety, transport, and administrative-tool companies with two to ten employees that cannot afford a dedicated public-sector sales role.
• Why now: With 22 problems from 18 agencies published at once, teams must repeatedly read and copy similar notices.
• First screen: Let a user enter a three-sentence description of their product, then show matching challenges alongside unmet application conditions.

2. Field-problem proposal maker

• What: Let public-agency staff record an inconvenient task in speech or a short note, then organise it into a problem description, current process, desired change, and trial criteria.
• Who: A frontline supervisor at a facilities-management corporation or an administration-team lead at a regional medical centre who deals with repeated complaints or manual reporting but has never written a digital-transformation project document.
• Why now: As agencies begin publishing problems first, it matters more to turn frontline pain into challenges that companies can understand.
• First screen: Include a field for the task that took the longest today and a field to attach the current document or photo.

3. Non-public-data negotiation helper

• What: Organise the needed data fields, period, update frequency, protection method, and alternatives, then create a draft agency discussion note and data-provision request.
• Who: Startups that need non-public data, such as hospital waiting times, road video, or facility-failure records, to test their products.
• Why now: Sixty-three companies applied to propose uses of non-public data, but actual provision still requires agency-by-agency decisions on rights and security.
• First screen: When a user enters the name of needed data, show four options: original data, data modified so people cannot be identified, aggregated values, and analysis only inside the agency.

4. A record book from trial results to procurement review

• What: Continuously collect meeting decisions, data changes, before-and-after trial figures, and field feedback, then turn them into a results report and operating plan.
• Who: A startup lead running a first field trial with a public agency and the agency contact who must review the following year’s purchasing budget based on its result.
• Why now: Trial funding of up to KRW 100 million is available, but purchasing, maintenance, and rollout to other agencies still need separate preparation after success.
• First screen: Ask users to enter three trial goals, the current figure, the target figure, and the person responsible for confirmation on each line.

A scene to watch for 30 minutes today

Share a screen with one startup chief executive who has applied for a public-agency challenge. Ask them to show, without skipping steps, the moment they found a recent notice through to the completion of their review memo. If they enter the same information more than once, or must confirm even one item needed to decide whether to apply through a separate call or email, a small test of a public-challenge matcher may be worthwhile.

Why this matters where you are

Public-sector challenge programmes may differ in their data-access rules, trial funding, and purchasing processes. You can check whether agencies in your market publish operational problems before issuing detailed product specifications, and whether trial records are later used in purchasing reviews. If those steps are handled in separate documents and systems, a tool that keeps the evidence connected may be useful.