The Pilot Graveyard:
Why Corporate Innovation Dies Before It Scales

A little over a year ago, I sat in on an innovation review at a large Brazilian industrial company. The CTO and the Supply Chain Director presented a brilliant pilot: an AI-powered control tower built to orchestrate logistics and distribution for one of Brazil’s largest telecom operators. The dashboards were flawless — designed with the support of a well-known strategy consultancy — and the early results looked promising.

A few months later, the project landed on the CIO’s desk for an assessment of integration, scalability, and information security. The verdict was blunt, almost surgical:

“There’s no way this works here — not even by a miracle.”

The control tower would require integrations with the ERP, TMS, WMS, and a host of other legacy systems, and it had been conceived on an architecture that couldn’t handle real-world data volumes, let alone meet the company’s cybersecurity and data-privacy requirements. The cost of turning the proof of concept into a pilot — and the pilot into an enterprise-grade application — made the project economically unviable. The PoC was flawless, but it was incompatible with the technology architecture of the applications already running in production.

This story isn’t fiction. And, more importantly, it isn’t the exception. It’s the rule.

A bitter truth nobody wants to talk about

McKinsey has tracked this phenomenon for years, and its research consistently shows that roughly 70% of digital transformation initiatives fail to deliver the results they promised. A study of more than 600 companies in 2022 found that only 20% captured more than three-quarters of their projected revenue gains, and just 17% achieved the cost savings they had expected.

Gartner goes further: it estimates that 80% of organizations that set out to scale digital business models simply failed to do so over the past several years. The reason cited most often isn’t a lack of technology. It’s a lack of governance and architecture to absorb it.

Bain & Company, in a 2024 analysis, arrived at an even bleaker number: 88% of corporate transformations fall short of their original goals. Not because the ideas were bad — but because the distance between the lab and real operations was never properly addressed.

In Brazil, the picture is just as telling. Major companies like Bradesco, Itaú, Vale, and a range of industrial players have poured billions into innovation ecosystems, digital hubs, and accelerator programs.

There are isolated wins. But the question rarely asked out loud is: how many of those initiatives actually reached operational scale?

Two executives, two agendas, one problem

To understand why this happens, you first have to understand who the protagonists of this quiet conflict are. And sometimes the conflict isn’t all that quiet!

The CTO (Chief Technology Officer) is, in essence, the explorer. Their mission is to identify new platforms, new architectures, and new applications — often sourced from startups whose solutions aren’t yet fully (or even minimally) secure, stable, and reliable — and to use them to enable new business models through technology. According to Deloitte research, 65% of CTOs name innovation as their number-one priority. It’s a role naturally oriented toward the future, toward experimentation, toward what doesn’t yet exist inside the company. The CTO tends to be the darling of the business units: pragmatic, solution-driven, able to show fast results in a controlled environment. And, let’s be honest, there’s a lot of hype around each initiative — which generates visibility and self-promotion for everyone involved.

The CIO (Chief Information Officer) operates on a different logic. Their responsibility is to keep the company running — today, tomorrow, and securely. Here, yes, security, stability, and reliability matter — and so, in turn, do governance, systems integration, compliance, critical infrastructure, technology risk management, and the annual SOX and information-security audits. A report by Harvard Business Review Analytic Services found that, in the eyes of many executives, the CIO still spends much of their time on cost control and IT crisis management — when they should, ideally, be co-creating business strategy.

To put it simply: the CTO explores the future with bold bets meant to change the business. The CIO has to make that future actually work while keeping the business running.

When these agendas converge, innovation scales. When they diverge — and in most non-tech companies they diverge systematically — you get what Silicon Valley has already nicknamed the ‘PoC Graveyard,’ the cemetery of proofs of concept.

The innovation that only works in PowerPoint or Figma

The problem has a well-defined anatomy. Labs, innovation teams, and R&D groups can demonstrate technologies that look revolutionary at small scale. A pilot with five users works. A controlled environment, free of the friction of real operations, delivers impressive results.

The picture changes completely, though, the moment that innovation has to be integrated into the company’s operations — with their decades of accumulated legacy systems, their regulatory compliance demands, their real transaction volumes, their SOX and information-security requirements, and the natural cultural resistance to change.

That’s when the CIO asks the question the CTO rarely asks during the pilot: does this work for the entire company? A solution that performs well for a hundred users can collapse with a thousand. A platform that runs smoothly in isolation can grind to a halt the moment it’s connected to even a handful of critical systems. An operating cost that looks reasonable in the pilot can be prohibitive at enterprise scale.

“A small innovation group can afford to look out over a three-, five-, or ten-year horizon. But when you actually try to take that step — to make the idea real and scale it across the whole company — that’s when the challenges show up.”

The observation belongs to Satya Jayadev, Vice President and CIO of Skyworks, and it captures the problem precisely. The challenge isn’t innovation itself. It’s the bridge between the lab and operations.

The cases that teach

In the early 2010s, Walmart invested heavily in digital innovation labs across the United States. Many of the initiatives were technically sophisticated. But only those deeply integrated into the company’s existing architecture — such as its logistics and analytics platforms — survived the test of operational scale. The rest were quietly discontinued.

GE, in building GE Digital, started from an ambitious vision: to turn the company into an industrial software powerhouse. The initiative produced genuinely relevant technologies. But it ran into severe obstacles integrating those solutions into the global ecosystem of a company operating across multiple, highly complex industrial sectors. Part of the strategy was “right-sized” — an elegant euphemism for saying that operational reality beat the lab’s vision.

In Brazil, the most emblematic case is Bradesco. inovabra (the bank’s innovation ecosystem) became a benchmark in the national financial sector. In 2022, the bank invested R$2.9 billion in the area, with more than 230 startups in the ecosystem. The results are real: contracts signed, solutions deployed, an innovation culture spread internally. But Fernando Freitas, then the bank’s head of innovation, was candid about the challenge: in one internal process, a project took 13 months just to graduate from the pilot stage. The growing push toward more mature startups — Series A and B — reflects exactly that lesson: innovating is easier than absorbing innovation into a structure with decades of accumulated critical systems.

The most revealing counterpoint comes from where you’d least expect it: Cirque du Soleil. The company is famous for pushing the limits of human creativity on stage. Backstage, it faced a different but equally complex challenge: an aging IT infrastructure, with decades of accumulated customizations and a technology team that shrank from 150 to just 13 people during the pandemic. Instead of launching innovation pilots disconnected from operations, Philippe Lalumière, the company’s VP of IT, did the opposite: he prioritized rebuilding the foundation first, migrating to a Cloud ERP in under two years with a two-phase MVP approach, and only then moving on to artificial intelligence initiatives. The result was an AI assistant for invoice management that cut handling time from 30 minutes to 2 minutes per query — already in production, with the potential to evolve into an autonomous agent. The sequence made all the difference in scaling innovation: first building a robust, simplified, and renewed foundation, and only then building innovation on top of that more modern architecture.

Amazon and Netflix tell the story that runs opposite to the PoC graveyard — and for a structural reason. Both were born digital and built their technology architectures with scalability as a premise, not as an afterthought. Netflix’s microservices culture and Amazon’s distributed-architecture principles, which gave rise to AWS, were no accident. They were architectural choices made before the problem existed.

This is what separates those who scale innovation from those who pile up failed pilots: a robust architecture, grounded in leading-edge technology, comes before innovation.

The CIO isn’t the villain — they’re the architect of reality

There’s a convenient narrative in organizations that casts the CIO as the executive who “puts the brakes on innovation.” In practice, the CIO is the executive who protects the company from poorly structured innovation — which can be even more expensive than not innovating at all, and can destabilize the business.

According to MIT Sloan research, 71% of CIOs already expect to play an active role in business strategy and innovation in the coming years. The role is no longer just guardian of business continuity. It’s that of a change engineer — an engineer who knows the terrain, who knows where the unstable foundations are, which systems can’t be touched without risk, and the invisible interdependencies that any lab pilot ignores by definition.

A Harvard Business Review survey found that 54% of CIOs believe IT strategy and business strategy are not sufficiently aligned in their organizations. Tellingly, 49% of CTOs admit that their technology initiatives aren’t adequately reflected in the company’s strategic objectives. Both sides recognize the problem. But they rarely build the bridge together — and before the pilot begins.

The real diagnosis: innovation without architecture doesn’t scale

The root of the problem isn’t technological. It’s structural and cultural.

McKinsey found that companies investing in cultural change are 5.3 times more likely to succeed in transformations than those that focus on technology alone. Culture, more than tooling, determines whether innovation becomes a product or becomes scrap.

In the specific context of the CTO–CIO conflict, that means innovation has to be designed for the real company, not for the ideal lab. Three concrete shifts make the difference:

• The CIO needs to enter the innovation process earlier, not later — not as a reviewer of what the CTO has built, but as the engineer of the real world’s constraints and possibilities. Architecture and operational scalability aren’t validation stages; they’re design variables. When that premise isn’t built in from the start of the pilot, it inevitably becomes the obstacle that kills the project.
• Innovation metrics need to include operational viability from the starting line. How many approved proofs of concept actually reach production? That conversion rate — from pilot to operation — is the metric that separates innovation programs that truly transform companies from those that collect hackathon trophies.
• And perhaps the most overlooked point of all: innovation needs an owner in the business, not just in technology. Initiatives confined to labs or IT teams tend to stall for lack of organizational traction. When business leaders define the expected value, the KPIs, and the operational adoption model from the outset, the odds of a pilot turning into a real solution rise substantially. Innovation without business sponsorship is an IT project. Innovation with business sponsorship is transformation.

Michael Porter famously said that operational effectiveness is not strategy. In the same way, an innovation lab is not transformation. Innovation that finds no path to real operations is just sophisticated cost.

The missing link: the bridger

Even when all three conditions above are in place — the CIO engaged from the start, conversion metrics established, and business sponsorship secured — there’s still a frequently invisible gap: who, in practice, builds and maintains the bridges between these worlds in the day-to-day of the project?

The March/April 2026 issue of Harvard Business Review offers a direct answer. In the article “Why Great Innovations Fail to Scale,” Linda Hill (a professor at Harvard Business School and one of the foremost authorities on innovation leadership) and her co-authors identify the single biggest factor distinguishing organizations that scale innovation from those that pile up pilots. It isn’t the technology. It isn’t the budget. It’s a specific kind of leadership they call bridging — the ability to collaborate effectively across organizational boundaries.

“No single team or company has all the capabilities, tools, or authority needed to move ideas from prototype to scale.”

Bridgers, as Hill defines them, are leaders with enough emotional and contextual intelligence to build mutual trust, translate between worlds that operate on different logics, and integrate the efforts of partners with frequently conflicting agendas. They don’t do this through hierarchical authority. They do it through relationships, credibility, and a rare ability to make each party feel that their interests are being taken into account.

In the context of the CTO–CIO conflict, the bridger is the executive who can sit down with the innovation team and the architecture team at once — and be taken seriously by both. They speak the language of the possible with the CTO and the language of the sustainable with the CIO. They neither dismiss operational constraints as bureaucracy nor write off bold ambitions as naïveté. They turn the tension between those perspectives into design.

The case of Delta Air Lines illustrates exactly what’s at stake. Nicole Jones, who built and led The Hangar — Delta’s first global innovation lab — faced precisely the problem of integrating innovation and operations. When her team developed a biometric boarding technology that required coordination among startups, internal IT teams, and federal agencies like the TSA, tensions were inevitable: Delta’s IT team had a deep aversion to risk after a costly outage the year before. Jones didn’t solve it with formal authority. She solved it by constantly reminding everyone of the shared ambition — improving the experience of millions of passengers — and by making explicit to each party how that connected to their own priorities. That year, Delta became the first major airline to take the stage at the Consumer Electronics Show.

What makes the concept of the bridger especially relevant for non-tech companies is that it doesn’t require creating a new box on the org chart. It’s a competency — and one that can be deliberately cultivated. Hill and her co-authors find that organizations placing bridgers in key innovation roles don’t just scale faster: they create an environment where trust among the teams that need to collaborate becomes a permanent asset, rather than a fragile achievement to be rebuilt with every new project.

In short: architecture solves the technical problem of scalability. Business sponsorship solves the problem of strategic relevance. And the bridger solves the human problem — getting people with different agendas, incentives, and languages to work together toward a goal that doesn’t yet exist.

The balance that makes the difference

The future of companies competing in an environment of intense technological change doesn’t depend solely on who leads technology. It depends on how these roles work together — and, above all, on when and how they start the conversation.

The CTO has to explore what’s possible. The CIO has to ensure that the possible is sustainable. The business has to ensure that what’s sustainable is also wanted — with clear targets, defined KPIs, and real commitment to adoption. And the bridger has to ensure that these three perspectives meet at the right moment, before organizational silos turn natural divergence into paralyzing conflict.

When this quadrilateral works, innovation stops being an expensive lab and becomes real competitive advantage. When it doesn’t — when the CTO builds in a vacuum, the CIO rejects at the end, the business takes no ownership, and no one is building the bridge — what should have been transformation becomes just one more line of cost on the balance sheet.

Buried in the silent graveyard of good ideas no one managed to scale.

---

Author’s note

I couldn’t fail to cite Linda A. Hill’s article in this issue of Harvard Business Review. Linda was my professor in the Harvard Program for Leadership Development (HPLP) — one of the most transformative experiences of my career — and coming across this issue at the airport on my way to SXSW 2026, while I was writing this very article, picked up precisely on its central theme. It was one of the happiest coincidences this project has given me. The concept of bridging that she and her co-authors develop in “Why Great Innovations Fail to Scale” is, in my view, one of the most urgent contributions to the debate on innovation leadership in recent years.

Sources

• HILL, Linda A.; TEDARDS, Emily; WILD, Jason. Why Great Innovations Fail to Scale. Harvard Business Review, Boston, Mar./Apr. 2026, pp. 74–85. [Adapted from the book Genius at Scale. Harvard Business Review Press, Mar. 2026.]
• McKINSEY & COMPANY. Losing from day one: Why even successful transformations fall short. McKinsey Quarterly, Dec. 2021.
• McKINSEY & COMPANY. The state of achieving digital value at scale. McKinsey Digital, 2022.
• GARTNER. Digital business scaling survey. Gartner Research, 2021.
• BAIN & COMPANY. Making transformation work: Lessons from 1,000+ programs. Bain Insights, 2024.
• DELOITTE. 2023 Global Technology Leadership Study. Deloitte Insights, 2023.
• MIT SLOAN MANAGEMENT REVIEW. The evolving role of the CIO. MIT Sloan Management Review, Cambridge, 2022.
• HARVARD BUSINESS REVIEW ANALYTIC SERVICES. Closing the IT-business strategy gap. Harvard Business Review, Boston, 2022.
• JAYADEV, Satya. Remarks on the challenge of scaling innovation. In: CIO Executive Council Forums, 2022.
• SAP SE; OXFORD ECONOMICS. The Value of AI with SAP. Oxford Economics, Oct. 2025. Available at: sap.com/documents. Accessed: Mar. 2026.
• SAP NEWS CENTER. Cirque du Soleil Entertainment Group implements AI-enabled invoice assistant. SAP SE, Walldorf, Oct. 2025. Available at: news.sap.com. Accessed: Mar. 2026.
• SAP APPHAUS NETWORK. Cirque du Soleil’s Leap into the Future through RISE with SAP. SAP AppHaus, 2025. Available at: apphaus.sap.com. Accessed: Mar. 2026.
• FREITAS, Fernando. Remarks on the innovation process at Bradesco/inovabra. In: Brazilian Corporate Innovation Forum, São Paulo, 2022.
• PORTER, Michael E. What is strategy? Harvard Business Review, Boston, Nov./Dec. 1996.