Getting AI initiatives up and running is getting easier for companies as experience in this relatively new domain ramps up worldwide. According to Gartner, “By the end of 2024, 75 percent of enterprises will shift from piloting to operationalizing AI, driving a 5x increase in streaming data and analytics infrastructures.” But don’t celebrate just yet, as success and scalability are often harder to come by. 

You still need to make sure you’re doing the necessary work to ensure your AI efforts will scale successfully. While a recent McKinsey Digital article states that, “Leading AI adopters (those that attribute 20 percent or more of their organizations’ earnings before interest and taxes to AI) are investing even more in AI in response to the pandemic and the ensuing acceleration of digital,” at the same time, many companies are failing to make the foundational shifts required to scale success. This includes defining new processes and implementing a modern data architecture. 

To help overcome these issues, the McKinsey Digital article offers five best practices for companies to break through the data architecture “gridlock,” including: 

  1. Take advantage of a road-tested blueprint. Fortunately, today’s data leaders can draw upon reference data architectures that can take the guesswork out of new deployments. Using a reference architecture helps reduce costs and enables a faster time to market. Instead of spending time on design, data leaders can build on a foundation over time to better meet their business needs.
  1. Build a minimum viable product—then scale. Rather than mapping out every intricate phase of a project, companies are encouraged to take a use case approach to AI initiatives, first building a “minimum viable product” and then adjusting it based on user feedback. This can include zeroing in on specific components spanning a wide range of use cases, including chatbots, virtual assistants, campaign reporting, and predictive analytics.
  1. Prepare your business for change. As data leaders look to modernize their company, it’s important to educate executives and business leaders on the value of new AI initiatives and the reasons to say goodbye to legacy technology. According to the article, “One telecom provider, for example, set up mandatory technology courses for its top 300 business managers to increase their data and technology literacy and facilitate decision making.”
  1. Build an agile data engineering organization. Not surprisingly, a well-performing, integrated team is critical to successful AI initiatives. Companies can achieve this by re-orienting the data organization to a product and platform model. Platform teams focus on building and operating the architecture, and product teams focus on developing business-driven AI uses cases.
  1. Automate deployment using DataOps. To help reduce project timelines, data leaders have started applying a DevOps approach to data operations. According to the article, “DataOps is structured into continuous integration and deployment phases with a focus on eliminating ‘low-value’ and automatable activities from engineers’ to-do lists and spanning the delivery life cycle across development, testing, deployment, and monitoring stages.” The process can remove time-consuming tasks for engineers and open up for time for coding.

If you’d like to learn more, read the complete McKinsey Digital article, which includes customer use cases and examples.