How to Balance Top-Down Decision-Making and Bottom-Up Innovation for Enterprise AI Adoption


Leaders today are under a lot of pressure—pressure to reduce costs, to drive revenue and, increasingly, to demonstrate how they are leveraging AI to achieve these goals. 

It’s not an easy ask. AI technologies are evolving rapidly. They’re learning and changing all the time, with new enterprise solutions, new models, new data and new use cases emerging daily. This makes agility an imperative as technical and non-technical leaders seek the highest-value use cases for the new technology within their companies.

At the same time, the sheer volume of decisions a leader has to make in this space can be overwhelming. Where should they apply their energy? Are they going to build or buy? Which tools should be deployed internally now to drive efficiency? Should they invest in partnerships, and what degree of investment in appropriate? And how do they prioritize AI tool adoption against other business goals?

To answer these questions and quickly move to adopt new technology across the enterprise, many companies lean on a top-down approach, with executives identifying the top areas the company should apply AI and telling their organization what is coming. Others take a hands-off approach, allowing for innovation, agility and employee- or team-determined need. But while most organizations choose one approach or the other, the real value comes from a balanced strategy that combines elements of both. 

A Top-down Approach Facilitates Enterprise Adoption 

A top-down AI adoption approach, which centers on having a governed model of deploying solutions more quickly, can help simplify this ecosystem. Leaders tend to use limited information available to select a set of use cases, and they create economies of scale by narrowing down the focused solutions and partnerships they will pursue in order to enable those use cases. This approach can allow for a perceived rapid response: solutions rolled out quickly, enterprise-wide with coordinated, visible efforts to drive adoption. Look at what we have enabled across all teams! Important considerations (such as building vs. buying, data privacy and security) can be considered centrally, enabling more control.  The unique sensitivities of each industry can be accommodated (for example, concerns around data types and use cases in healthcare or financial services), and executives are able to choose the solution best-suited to their risk profile. 

This type of strategy focuses on minimizing exposure to risk and assumes a clear understanding of the value proposition. While companies that take an exclusively top-down path appear to respond quickly, they often observe low adoption for expensive solutions selected and don’t actually realize the value anticipated in cost savings or efficiency. Why? The use cases that were anticipated by senior leaders, with solutions crafted centrally, weren’t actually the highest-value applications for AI. Sunk costs pile up quickly, and expensive pivots are considered. 

A Bottom-up Approach Can Help Uncover New Use Cases 

The bottom-up approach, in contrast, relies on grassroots innovation to surface use cases for AI. With this approach, leaders either empower employees to weave AI into their daily work as they desire—or leave them alone to do so.  Employees uncover custom use cases that might otherwise have never been envisioned by leaders looking at it from a high level, and they bring their own AI solutions to work, moving along with the market to test available technologies against real business challenges. 

But without a top-down mandate, what motivates employees to use AI? Employees have discovered on their own what research by Boston Consulting Group and Harvard Business School formally reported: Using AI makes knowledge workers significantly more productive—they completed 12.2% more tasks on average and completed tasks 25.1% more quickly and produced 40% higher quality compared to a control group. Employees who used AI at work also report that their jobs are easier and more enjoyable.

In fact, most employees are already using AI on the job—75% of them, according to recent research by Microsoft and LinkedIn. And more than three-quarters of those who do so are using their own tools, not company-provided ones. Where companies don’t actively promote AI use, more than half of the employees surveyed report that they’re hesitant to reveal that they’re applying AI to their most important tasks. They’re worried they’re going to get in trouble or put their jobs at risk.  

So, if companies can reap efficiency benefits without a centralized engine, why consider any other way? The problem with a completely unguided approach is multi-faceted: (1) Companies can’t amplify the efficiencies that select employees discover, (2) companies lose important control over privacy and security risks, and (3) companies end up with an expensive web of disparate solutions for similar use cases. This is also not ideal, which implores companies to consider an alternate hybrid approach. 

Creating an Engine to Balance Both Approaches

It’s not a binary choice between driving decisions from the top or through grassroots innovation—true transformation requires both. Here’s how you can lay the foundation to balance experimentation with a centralized engine to execute on the highest-value use cases:

  • Create a clear North Star to ensure that your organization’s values and goals will guide decision-making when it comes to AI. Every leader should have clarity on where they need to innovate, where the company is going and what biggest roadblocks and risks are. 
  • Spend time to roughly identify the highest-value use cases—where new technology could move the needle in your organization. This could be reducing time to production, improving quality, driving active users, amplifying subscriptions or any other business goal. By prioritizing based on value and associated risk profile, you can find the areas where investing in innovation will be most fruitful.
  • Within a zone of high-value use, create an environment where safe experimentation is promoted and valued. This can be done by setting security and privacy guardrails, allocating a defined budget for experimentation and, most importantly, communicating your desire for teams to innovate. Encourage teams to test many solutions before deciding on a path forward.
  • Continuously collect data on what’s being tried, and what’s working. Use traditional operational metrics to measure the impact of AI innovation on your goals—unless you’re an AI company, your business goals should not be materially changed by a new technology; rather, you should use the technology to further differentiate your company against its competition. 
  • Cultivate learning within your organization to stimulate cross-functional innovation. A Center of Excellence can serve as a hub that sources ideas from employees and links them to dedicated central investments and rigorous decision-making around solutions. 
  • Be ready and willing to invest in and proliferate ideas that show a proven track record of success. Once the experimentation has proven value creation, don’t waste time in doubling down on the solutions that work. At the same time, encourage teams to re-evaluate as technology evolves and changes. 

Leveraging grassroots efforts to prioritize the highest value-use cases and harnessing corporate-level horsepower to set appropriate guardrails for innovation can ensure that you maximize the benefits of both. In the end, successful enterprise AI adoption at scale has more to do with culture, positioning and change management than with the technologies involved. In fact, while every organization is different, executives should expect to dedicate the lion’s share of AI efforts to business and people transformation.  

If you invest the appropriate time and resources to create an effective business-driven AI innovation engine, you can sidestep the common sense of overwhelm and be confident that you’re harnessing the innovative power of your organization to reap the biggest value from new AI technology—now and in the future. 

About the Author

Molly Lebowitz, Senior Director, Propeller. A strategic leader, practiced engineer, and critical thinker, Molly Lebowitz has extensive experience helping technology organizations tackle large-scale, complex operational challenges and transformations. From operational excellence to market intelligence, strategic planning, and executive-level decision-making, Molly is adept at helping leaders in the tech industry energize, reconfigure and up-level their teams and business. Her experience in software, hardware, media, and online travel brings the expertise and perspective to drive transformative results. She holds a bachelor’s degree in engineering from Cornell University. 

Sign up for the free insideAI News newsletter.

Join us on Twitter: https://twitter.com/InsideBigData1

Join us on LinkedIn: https://www.linkedin.com/company/insideainews/

Join us on Facebook: https://www.facebook.com/insideAINEWSNOW





Source link