top of page

Corporate-Startup Partnerships: The POC Passed. Now What?

8 hours ago
5 min read
A technician holding a digital caliper, measuring the dimensions of a metal part
Passing the measurement check is only step one. Once the pilot clears, the real hurdle is who signs the purchase order. (Photo: Hans Westbeek / Unsplash)

Innovation leads at large companies often ask me: "Over the past two years we've run quite a few POCs, and the startups have all worked hard. So why does almost nothing come of them?" Every time I hear this, I ask a question back: Before the POC started, did you agree on who would pay for it if it succeeded?


That question is the most overlooked part of corporate startup engagement (CSE). Everyone pours energy into finding startups, running events, and launching pilots, but very few design for what happens after the pilot. Today I've pulled together three recently published international cases and compared them with two new sets of survey data, to see how others have gone all the way from POC to contract.


Today's Highlights: Three corporate-startup partnerships that went from pilot to contract


Mitsubishi Electric: from an innovation challenge to a signed contract in nine months


In its corporate-startup partnership survey report published in October, Plug and Play documented the collaboration between Mitsubishi Electric and Facilis, an AI visual inspection startup. Plug and Play connected the two through the Mitsubishi Electric Innovation Challenge. The pilot kicked off in January 2025, results were reviewed in May, deployment began in August, and a commercial contract was signed in September, with the scope expanding to multiple business units.


The reported results: detection accuracy above 99%, a 15% reduction in defects, and inference time under 60 milliseconds per image. What I find even more noteworthy is the timeline: about nine months from pilot launch to contract, with a clear review checkpoint at every stage, so the pilot was never allowed to drag on indefinitely.


Kajima Corporation: define the pain point first, then go find a partner


The second case in the same report is Kajima Corporation, the Japanese construction group. In August 2023, Kajima put forward a very specific need: its job sites had piled up huge amounts of video that was hard to search and analyze. The following month, Plug and Play matched it with the startup Archetype AI. The two ran a POC on this exact problem, and Kajima went on to become Archetype AI's first paying design partner.


The published results include analyzing 11,957 videos, identifying 662 target clips, and finding relevant footage in under 60 seconds. The key was that the need was defined to a measurable level, so once the POC wrapped up, success or failure was obvious at a glance.


Globe: a year-round matching platform that goes after one real order first


The third case is in Asia. 917Ventures, the venture builder under Philippine telecom company Globe, launched its Velocity program in 2025. It runs year-round with no cohorts, letting startups run pilots, POCs, and technical integrations across the corporate network of Globe and the Ayala Group. According to a mid-November report by local tech media, the program has drawn more than 500 startups onto its list so far.


Four startups are currently advancing commercial partnerships. Among them, behavioral AI platform Netopia AI signed the program's first commercial contract with digital advertising company AdSpark in August 2026, while the other three are still in talks. 917Ventures has also set a goal of closing 10 startup deals in 2027. More than 500 startups on the list, and so far just one public commercial contract: that is what the corporate-startup engagement funnel really looks like.


By the numbers: everyone gets stuck after the POC


Plug and Play surveyed 203 startup leaders. About 73% had done a pilot or POC, but more than 86% said slow corporate decision-making and procurement was a major challenge. About 58% said it takes 3 to 12 months to reach a formal agreement, and 14% said it takes more than a year. 41% named revenue as the most important measure of a successful partnership, while only 4% chose "completing the pilot." The most common reason partnerships stalled was a shift in the corporation's strategic priorities (25%), and another 13% ended because the internal champion left the company.


A report released on November 16 by the Federation of Korean Industries (FKI) shows the same gap: among 58 startups with open innovation experience, 56.9% had done a POC, only 25.9% reached commercialization, and just 10.3% were able to scale.


Jesse's Take: The POC isn't the finish line. The purchase order is.


All three cases are success stories published by the corporations or accelerators themselves, so there's selection bias; the deals that went nowhere usually don't get written up. But precisely because they are success stories, what they got right is worth comparing against. Large companies in Taiwan tend to lean toward looking inward, with slow, hierarchical decision-making and long procurement and legal processes, while startups have limited resources and worry about doing work for nothing. These breaking points are only more pronounced in Taiwan.


First, frame the challenge precisely enough to measure


Kajima's need wasn't "adopt AI"; it was "make job-site video easy to find and understand." This is what I often call Define, the first step of DHEA: define the problem clearly, then propose a hypothesis you can test.


The more precise the need, the less the startup has to guess, and the less likely the company is to discover only after the POC that "this isn't what we wanted."


Second, map out the path to a contract before the POC begins


The Mitsubishi Electric case had clear checkpoints for review, deployment, and contract signing. Before a POC kicks off, I recommend writing down the KPIs and benchmarks for success, plus which business unit will buy, and from which budget, once the targets are hit. Failed tests have value too, but only if both sides agreed in advance on what success looks like.


Third, the champion needs to sit in a business unit, not just the innovation team


In the survey, 13% of partnerships ended because the champion left, and 25% were called off because of a strategic shift.


In Taiwan, what I see most often is an innovation department that runs beautiful events while no one in the business units picks up the ball. You need at least one business unit head who is willing to put their name on it and has budget in hand, so the partnership has somewhere to land.


A first step you can take next week


List every POC you're currently running or have just wrapped up in a single table, and fill in only three columns for each:


  1. What are the KPIs for success?

  2. Who signs the purchase order once the targets are met?

  3. Where does the budget sit?


If any of the three columns is blank, that project is likely to become the next one stuck at the POC stage. There's still time to fill in the gap now.


Further Reading


Related posts on JCXO


Original sources

Comments


©2025 | Omni Optimize Your Biz | JCXO Consultancy | JC Capital

Thanks for subscribing!

bottom of page