Open Innovation Challenge: How to Run a Startup Challenge That Ends in Pilots

In September 2024, Airports for Innovation (A4I), a network of airports, opened a global call for startups in four categories, from sustainable aviation to smart luggage. By the end of October, 292 proposals from 283 startups in 36 countries had arrived. Twenty-five finalists pitched in January 2025, and in May five winners started pilots at partner airports, including a trial of vertical solar towers at Munich Airport (Munich Airport press release, May 2025).

That is about one pilot for every 58 proposals, and it took eight months. The example shows the time and selection involved in moving from an open call to tests with operational teams.

If you run corporate innovation or an accelerator programme, you may know the other version. The call gets attention, the final gets photos, and three months later the operations director whose problem it was still has no pilot in his calendar. This guide shows how to run a challenge that ends differently. It covers the problem statement, the pilot commitment, the open call, the shortlist, the clarification meetings and the measures that matter.

For a challenge intended to lead to pilots, agree the operational owner, test conditions and route to a contract before opening the call.

What is an open innovation challenge?

Start with the definition, because the term covers several formats. An open innovation challenge is a public call in which a corporate, an accelerator or a public body publishes a specific business problem, invites startups and other external solvers to propose a solution, selects the best fits and tests them in a paid pilot or proof of concept (POC). The call usually runs for several weeks, and the selection takes another one to two months.

Three neighbouring formats are easy to confuse with it, and each answers a different question:

The key difference is where the solution comes from. In a hackathon it is built over a weekend, which is why our guide on how to host a hackathon worries most about the Monday after. In a challenge the startup brings a working product, so the risk shifts to selection and integration. A challenge also borrows from the other two formats. If its final ends with a jury vote, the scoring and prize questions in our guide on how to run a pitch competition apply. And if you want to know why the corporate in the room behaves like a buyer rather than a mentor, our guide on venture clienting explains the model behind it.

Why so many challenges end with a winner and no pilot

Whatever the format, the demand exists on both sides of the table. In the Collaborate to Innovate 2025 survey by the German Startup Association and Accenture, around six in ten established companies said they find it hard to identify suitable startup partners. Eighty per cent preferred to meet candidates in person at events or conferences (more than 500 respondents, reported by Startbase and Unternehmeredition, August 2025). A challenge solves the first problem, because it brings candidates to you. Yet only 28 per cent of the companies in the same survey saw high returns from their cooperation projects.

The findings point to challenges in both finding suitable startups and turning cooperation into useful results.

The long-run ratio after the pilot looks similar. STARTUP AUTOBAHN, the Stuttgart innovation platform run by Plug and Play with automotive partners, reported 320 pilot projects with more than 220 startups between 2016 and early 2020, and more than 60 of them implemented (organiser press release, February 2020). Roughly one pilot in five made it into operation. If the pilot is the narrowest part of the funnel, the whole challenge has to be built around it.

Three gaps explain most of the drop between the final and the pilot:

• No owner: the innovation team ran the challenge, but the operations director whose budget pays for the pilot never committed to it.
• No terms: budget, test site, data access and the position on intellectual property (IP) were left for after the final, when every week of delay costs momentum.
• No contact: the shortlisted startups never spoke to the people who know the problem, so their pilot proposals were built on guesses.

Assign a business owner who can take the selected proposal through the next approval and delivery steps.

Six stages from problem to pilot

Each of those gaps can be closed before the first startup applies. The sequence below follows the timelines that two public challenge programmes in Singapore publish: the Open Innovation Platform of the Infocomm Media Development Authority (IMDA) and the Aerospace Open Innovation Challenge 2026 of Enterprise Singapore.

1. Write the problem, not the solution

The problem statement decides who applies. Jeppesen and Lakhani studied 166 science challenges and more than 12,000 solvers on a broadcast search platform. Solvers whose expertise lay further from the problem’s field were more likely to submit a winning solution (Organization Science, 2010). The best answer often comes from a field you did not think of.

That only works if the statement describes the problem rather than the technology you expect. “We need to know within ten minutes when a refrigerated shipment leaves its temperature range” invites sensor makers, software firms and packaging startups. “We are looking for a tracking platform” invites only platform vendors. Add the current situation, the constraints (systems, regulation, sites) and what a successful pilot would show after three months.

2. Commit the pilot before the call opens

Make the pilot commitment a prerequisite for launching a challenge that promises operational testing. Before anything is published, the operations director who owns the problem should sign a one-page pilot commitment: the budget range, the test site, which data the startup will get and who on his team will spend time on it.

The IP position belongs on the same page, because good startups read the terms before they apply. IMDA sets a clear default for its POC and prototype challenges: all IP created from the solution provider’s work on the challenge belongs to the solution provider. Problem owners may propose IP and exclusivity terms before development begins (IMDA Open Innovation Platform FAQ). Whatever you choose, publish it with the call.

If the commitment is not yet approved, resolve the open points before inviting startups to spend time on a pilot proposal.

Prize money can help, but it is not the main draw. On the IMDA platform, problem owners set the prize, typically 30,000 to 50,000 Singapore dollars, and around 30 per cent goes to the finalists to cover development costs. Explain what a paid pilot offers beyond prize money, including access to users and the possibility of a reference customer.

3. Run the open call

Once the terms are agreed, prepare an application process proportionate to the selection task. IMDA runs an eight-week application phase. Keep the application short: Enterprise Singapore asks for a participation form and a deck of five to ten slides, and DHL asked for a video of up to five minutes for its 2025 Fast Forward Challenge in the Middle East and Africa. Require a working product or at least a minimum viable product (MVP), as DHL did, so applicants understand the level of product readiness needed for testing.

Promote the call where startups already look: accelerator networks, startup conferences, university spin-off programmes and the communities of your existing partners.

4. Shortlist with the problem owner in the room

The innovation team can sort out ineligible applications, but the shortlist belongs to the people who will use the solution. At DHL, applications were reviewed by potential end users and subject matter experts. The finalists were then judged on five published criteria, which make a good starting point:

• Innovation: what the solution does that current approaches cannot.
• Solution market fit: how closely it matches the published problem.
• Commercial viability: whether the business model will survive past the pilot.
• Solution readiness: whether the product works today, with references.
• Sustainability: the environmental and social effect, where it matters to the problem.

Add one criterion of your own: the integration path. A startup that cannot connect to your systems in a few weeks will spend the pilot on interfaces.

5. Hold clarification meetings before the final

Use the shortlist phase to clarify the operational questions that applications cannot fully answer. Enterprise Singapore schedules one-to-one clarification sessions between shortlisted innovators and the companies that own each challenge statement. In 2026 they ran for about five weeks, from late May to the end of June, before the final selection in mid-July.

These meetings change the quality of the pilot proposals. Picture a 30-minute call in which the logistics lead mentions that the loading dock has no wireless network. Knowing that constraint early lets the startup assess whether and how its product could be tested at the site. The startup learns about data, systems and constraints; the problem owner learns whether the team can deliver.

Plan them like any other meeting programme: fixed time slots, a cap per problem owner and a record of who met whom.

Converve schedules these one-to-one meetings in time slots you define, with rules for who can meet whom and a limit on meetings per problem owner. Shortlisted startups and business unit owners request and confirm meetings in a shared schedule, and every meeting is logged, so you can later trace which conversation led to which pilot. Our solution for startup and investor events shows how the same meeting matrix runs founder and corporate programmes side by side.

6. Decide fast and start the pilot

The final selection should come with a contract template that is already approved. IMDA plans three to six months for the POC or prototype phase itself. A4I started its pilot phase with a kick-off event on the same day it announced the winners.

Give the selected team a named contact, agreed next steps and a start date so the pilot can move into delivery.

Where the event fits

Of the six stages, only the final is public, and organisers tend to treat it as a show. A better approach is to treat it as a meeting day. DHL held the final of its 2025 Fast Forward Challenge as an invite-only pitch event in Abu Dhabi. A4I invited its 25 finalists to pitch in person.

If you run a startup conference that hosts a corporate’s final, plan the day in two halves. In the morning, the finalists pitch. In the afternoon, each finalist meets the business units that might pilot its product, in pre-scheduled 1:1 slots. The audience learns about the problem, and the problem owner gets the conversations that lead to a decision.

The public launch can be an event too. A reverse pitch, in which the corporate presents its problems to a room of startups, is a good way to open the call. Our overview of pitch event formats explains how it differs from a competition.

How to measure an open innovation challenge

The operations director will ask one question at the end: did it fix the problem? Track progress from eligible applications through testing and implementation:

• Eligible proposals per challenge statement, not total applications.
• Share of shortlisted startups that had at least one clarification meeting with the problem owner.
• Pilots started within 60 days of the final selection.
• Days from the close of the call to the first pilot start.
• Pilots that turned into a contract or rollout after six to twelve months.
• Startups that apply again to your next challenge.

The last figure is a good test of your reputation. A startup that spent weeks on a proposal and heard nothing will not apply again.

Not every problem needs an external supplier, though. When the answer probably sits in data the business already holds, a datathon with internal and external analysts can be the faster first step, and our guide on how to organise a datathon explains how to clear that data before anyone works with it.

Frequently Asked Questions

What is an innovation challenge?

An innovation challenge is a public call in which a corporate, an accelerator or a public body publishes a specific business problem and invites startups or other external solvers to propose a solution. The organiser selects the best fits and tests them in a paid pilot or proof of concept. The A4I airport network, for example, received 292 proposals for four challenge categories and started five pilots (Munich Airport press release, May 2025).

How do you run an innovation challenge?

Run it in six stages: write the problem statement, commit the pilot terms with the problem owner, run an open call of about eight weeks, shortlist with the future users of the solution, hold one-to-one clarification meetings with the shortlisted startups, then decide and start the pilot. The second stage matters most, because a challenge without an owner and a budget rarely ends in a pilot.

How long does an open innovation challenge take?

Plan for four to eight months from the start of the call to the start of the pilot. IMDA in Singapore uses an eight-week application phase, four to eight weeks for evaluation and finalist selection, and three to six months for the proof of concept itself (IMDA Open Innovation Platform FAQ). The A4I call opened in September 2024 and its pilots started in May 2025.

Who owns the intellectual property in an open innovation challenge?

It depends on the challenge terms, so publish them with the call. A common default is that the startup keeps the IP it creates. IMDA states that all IP created from the solution provider’s work on a challenge belongs to the solution provider, while problem owners may propose IP and exclusivity terms before development begins. This is general information, not legal advice; check the terms with your legal team.

What is the difference between an innovation challenge and a hackathon?

In a hackathon, teams build a prototype on site within 24 to 72 hours. In an innovation challenge, startups apply over several weeks with a product that already works, and the result is a pilot rather than a prototype. A hackathon is a good way to explore a problem; a challenge is a good way to find a supplier for it.

Can you give me some ideas for an innovation challenge?

Good challenge topics are specific problems with an owner and a measurable result. Examples from published challenges include sustainable aviation and smart luggage handling at airports (A4I, 2024), and life sciences supply chains and outdoor autonomous logistics (DHL Fast Forward Challenge, 2025). Choose a problem your business unit would pay to fix within a year.

Conclusion: commit the pilot before you open the call

Finding suitable startups and converting their proposals into pilots both need preparation. Agree the problem, test conditions and operational responsibilities before the call goes out. Write the problem rather than the solution. Get the operations director to sign the pilot terms. Let the future users shortlist, and give every shortlisted startup a conversation with the problem owner before the final.

Use the final selection to confirm the next delivery steps, then report pilots started and subsequent implementation alongside application numbers.

If you want to schedule the clarification meetings and the finalist day as one matched meeting programme, get in touch with Converve.