SEVEN ESSENTIAL STEPS TO DRIVE A SUCCESSFUL QUANTUM COMPUTING PROOF-OF-CONCEPT PROJECT
Key players insights
Digital disruption prompted a wave of transformations in recent years as businesses moved to address customer expectations for better, faster experiences. Today, digital transformation is a central tenet of every modern business strategy. Quantum computing is now arriving on the scene to drive digital transformation even further. As I mentioned in a recent Forbes article, quantum computing will have a tectonic impact on many industries and solve what used to be unsolvable challenges.
Boston Consulting Group (BCG) says that quantum computing could create $450 billion to $850 billion in value in the next 15 to 30 years.
That’s why tech giants like Amazon, Google and IBM are investing heavily in quantum computers, and it’s why leaders in finance, health care, logistics, and other sectors and disciplines are exploring quantum. These explorations often take the form of a proof of concept designed to test the waters, build internal expertise and avoid falling behind competitors. Based on our experience, here are seven essential steps organizations can take to make their quantum proof of concepts a success.
- GAIN EXECUTIVE SUPPORT
CTOs and CIOs typically don’t run quantum teams directly. They often don’t have the time to do that, and they may not have the deep quantum expertise. Instead, companies appoint team leads, often a Ph.D. with quantum computing knowledge and some management experience.
Initially, these quantum computing teams are asked to survey the corporate landscape, identify problems that are candidates for quantum solutions and prototype some solutions. These prototypes do not yet contribute to the bottom line, yet might cost tens of thousands of dollars.
Ensure that executives are aware of these proof-of-concept (POC) projects. Work to align executive and quantum teams on why you are exploring quantum computing. Ask and answer: What do we hope to gain? Who are our partners? Are we exposing any secrets? What’s next?
- BE REALISTIC
Breakthrough algorithms take time to develop, and those breakthroughs usually don’t happen via a POC.
You are not going to create a magical invention in three months. Instead, you are going to use that time and effort to do a test case. Use the data you collect from your POC to extrapolate what will be possible for you to do in six months or two years as quantum hardware improves and your algorithms mature.
What we often see is that organizations identify an academic article that describes a quantum algorithm in their field, use the POC to implement this algorithm on their data and then to experiment with some tweaks. This is a reasonable approach, which builds trust in the field and its importance to the company.
- THINK ABOUT NEXT STEPS
Let’s assume that the POC is a success. Don’t just celebrate and put the project to sleep. Outline what your organization should do to move the effort into production. Leverage your success to build the quantum team and get additional required resources to continue moving the effort forward. Take a crawl-walk-run approach to quantum. Plan to succeed, and then move to what’s next.
- UNDERSTAND THE LIMITS OF TODAY’S QUANTUM COMPUTERS
Today’s quantum computers may not offer the power that you need to solve your problem right now. But be aware that quantum computers are on a rapid path of improvement. It’s somewhat akin to Moore’s Law, which was the driving force behind rapid development of faster and denser classical computers. The same basic force applies to quantum computing. But some people — such as Hartmut Neven of Google’s Quantum AI Labs — say that quantum is actually moving faster these days.
- INVOLVE DOMAIN EXPERTS
For quantum to gain business traction, it cannot remain within the realm of quantum enthusiasts. Pull in sector-specific domain experts and leaders from company verticals, as well as broader disciplines such as supply chain, operations and IT. Get them excited about what quantum computing can do to make them more successful and to enhance their value for the company. If you’ve got executive support and domain expert buy-in, you will be able to move faster.
- CONSIDER THE RUNTIME ENVIRONMENT
There may be a technical question about the runtime production environment. But for a POC, be aware that you often don’t need to worry about that. Indeed, you might choose to start your POC today and leverage the POC to test out several runtime environments. During the POC, try to identify the platform you ultimately want to use.
- MAKE SURE YOU HAVE THE RIGHT PARTNER
Select a partner that has done a quantum computing project before and has deep quantum expertise. That way, the partner will be able to apply its experience and may be able to leverage its previous learnings for you.
Choose a partner whose business goals are aligned with yours. Consulting organizations sell hours, so they might not be highly motivated to teach you how to do things yourself. When you partner with a product company, the best outcome for both sides is that you will become an expert yourself.
Why am I qualified to suggest these steps? Because we have done it before. For instance, we worked with a multinational financial services firm to determine how many qubits it would need to get the accuracy it wants for option pricing. Now the company can gauge when that level of computing power will be available and understands its option pricing goals are within reach. With a leading aerospace company, we developed a quantum algorithm for advanced object trajectory tracking. This was a hybrid algorithm, meaning the algorithm runs partially on quantum computers and partially on classical computers to overcome the existing limitations of quantum machines.
Quantum computing may seem like a daunting challenge. But with the right partner and approach, it’s not that scary. You could complete a proof of concept in two or three months and without a big investment. And there’s a big payoff: building your team and arsenal of quantum algorithms, and putting you ahead of your competitors in the long run.