In an office somewhere, leadership continues to discuss just how effective the cloud would be for collaboration and efficiency but hesitates to make a move toward an integrated system. Elsewhere, another company spends thousands on blockchain research and development with little to show for it. And then there’s that artificial intelligence and machine learning team that keeps sharing news of great outcomes but doesn’t have anything ready for production just yet.
Does this sound like something you’ve seen or dealt with at work?
There are always challenges when attempting to employ emerging technologies, especially when there’s no clear pathway toward adoption.
Unfortunately, it can be easier to get caught up in the hype than to make a robust road map.
How New Technologies Get on Board
New technologies typically infiltrate companies in one of a few ways. Maybe an executive read about the success a competitor achieved with the application of a new piece of technology, or a senior leader attended a conference and saw an opportunity to create competitive advantage by using the latest, greatest tool. Perhaps an article or a podcast started it all.
Other times, though, there’s an almost grassroots effort to incorporate emerging technology into existing workflows. Maybe vocal members on a team somewhere within the company are actively exploring an area of interest, and they pitch leadership on why it would be a game changer for the company. Another person might want to transition careers into a new domain, so he or she starts driving adoption to accelerate that change through a proof of concept.
Both entry points often begin with the construction of prototypes designed to demonstrate how some new technology could work. These one-off solutions are typically built in a silo, meaning they have little executive commitment and are free to ignore the burden of scalability.
After all, prototypes are easy to create, inexpensive, and they don’t come with any long-term commitments. The due diligence necessary for building production-ready systems simply isn’t needed during this initial stage. This makes for quick internal wins that executives can applaud in celebrating the company’s culture of innovation. Not to mention that this exercise is fun and exciting for employees and offers a brief distraction from the daily grind. Then, after the excitement dissipates, the projects sit in stasis.
These scenarios reflect the traps companies have fallen into for decades.
Failure to Launch
The latest technologies to sweep the corporate world — AI and machine learning — face the same kind of adoption challenges, and they provide helpful context when examining the “death by prototype” problem.
These technologies represent a black box to many business leaders. This means executives are often hesitant to move past the prototype phase, even when initial implementations yield positive results.
With these types of projects, it’s fairly easy to show success in a controlled environment. When AI and machine learning fail, though, the results can be devastating. Understandably, risking business operations by depending on the technology at scale is an uncomfortable prospect for many leaders.
This also presents a unique psychological hurdle: A successful AI or machine learning implementation will almost certainly eliminate work previously conducted by humans. This might excite executives who are generally optimistic about an automated future, but a 2019 Edelman survey shows that only 30% of the general public has a positive outlook regarding AI in the workplace. It’s not hard to see why AI projects fail. Who wants to actively push for technology that they believe could make their roles obsolete?
Then, of course, there are AI’s programming challenges. Finding and recruiting the personnel with the skills to lead and execute a successful, large-scale implementation takes capital. Most organizations simply aren’t ready to match the faith in an outcome with the investment required to realize it.
However, those that have taken that leap of faith are often rewarded handsomely. When surveyed by McKinsey & Company, a combined 78% of businesses said that they were capturing either significant or moderate value from AI adoption across different business areas. A mere 1% noted “no or negative value” after implementation.
Although technologies such as machine learning and AI challenge the human psyche in ways that few others can, they promise to give bold business leaders tangible competitive advantage. If you hope for the same results within your business, here are five recommendations to consider before taking the prototyping leap:
1. Consider finances first.
Focus your prototype investments in the areas of the business that have the financial potential to scale. If that capital isn’t already on hand, map out a comprehensive plan for securing it. If you know you’d have trouble securing the resources to move beyond proof of concept for a prototype, avoid investing too much time and energy into it.
2. Identify the problem.
Be clear about what you’re trying to accomplish by deploying AI, machine learning, or any other functional technology solution. The value derived from the implementation must be scalable enough to justify the investment, so focus on areas of the business where you can achieve this. If your prototype will only generate nominal incremental value or if there’s no line of sight to a return on investment, look for areas that could benefit more from prototyped solutions.
Finally, not all prototypes must lead to revenue. Understand the line between prototypes that provide new layers of value by expanding knowledge and ones that can successfully operationalize.
3. Create a full-fledged plan.
Don’t begin building a prototype without first discussing and planning — at a high level — what comes after the prototype stage. Having general consensus and buy-in on what follows a successful prototype will accelerate the maturation stages of the project and keep leadership in tune with the longer-term goals of the prototype’s initiative. Plotting out these overarching next steps (and spending the hours to do so) will keep teams accountable for moving past the prototype stage.
4. Build a prototype portfolio.
If you’re committed to prototyping AI, machine learning, or other solutions, building and managing a portfolio of prototypes will unify your business’s experimentation with the technologies while maximizing knowledge sharing. Besides this, prototype portfolios will drive superior outcomes. When you compare prototypes against one another, you’ll advance your ability to identify and define success.
5. Develop a promotion program.
A successful prototype promotion doesn’t guarantee operationalization. You might have multiple levels of prototypes that are in different stages of their life cycles. For example, you could outline a tiered approach where the first tier confirms the possibility for value, the next tier tests scaling approaches, the next tests operationalization, and the final tier yields a production-ready solution.
Beginning any technological implementation without a plan for capturing real value could be a recipe for failure.
A prototype is a great step in the right direction, but simply building a prototype should never be your end goal.
AI, machine learning, and other emerging technologies hold unbelievable promise. But the companies that succeed with these solutions have to take bold risks, overcome implementation challenges, and ultimately, build more than just a prototype.