Continuous Engagement and Innovation

In the digital world, formerly discrete activities have become continuous. In the old world of IT, software deliveries were considered discrete events. We created projects to deliver software capabilities, set requirements for them, deployed code, and then moved on to the next project. At least that was the theory. We knew those IT systems would later need enhancements, bug fixes, and something odd we called “maintenance.” Odd because software doesn’t need maintenance. Unlike a car that needs oil or filter changes, software keeps doing the same thing forever. We “maintain” software mostly to change what it does.

These maintenance activities were considered an accidental consequence, an unfortunate need to keep spending money on something already “finished.”

Our outlook is very different in the digital world. We recognize that a company’s composition and goals constantly change, its market environment changes, opportunities present themselves, competitors act competitively, pandemics infect people, wars disrupt supply chains—well, since software is critical to what companies do today, and what they do constantly changes, obviously the software will need to change. More than that: we want it to change so that we can compete better, remain resilient—and not go out of business.

So rather than discrete deliveries, modern software practice is about continuous change. IT technique reflects this: we practice continuous integration, continuous deployment, and continuous delivery. Machine learning models are adjusted with new learning. We use the cloud to constantly adjust our infrastructure and other IT resources. DevOps teams own not only creating their products but also operating and adjusting them over time.

Whether we’ve noticed it or not, we are also in a world of continuous, rather than discrete, engagement. IT organizations used to engage with lines of business when planning a project and again when they delivered the resulting IT system. As agile techniques became common, periodic engagement was replaced with continuous engagement (sometimes mediated by a product owner or a Scrum Master).

We also continuously engage with customers by bombarding them with “How did we do?” emails and net promoter score (NPS) queries, or more subtly, through data and analytics and monitoring. Game designers can release new game features and then observe how players react. Designers can A/B test features and use the results to inform their designs. Perhaps the longer movement from memos to emails to Slack channels and text messages is another sign of this move to continuous rather than discrete engagement.

Interestingly, today’s challenging areas are often those where engagement has not yet moved from discrete to continuous. For example, CIOs and CFOs only tend to engage periodically, especially around budgeting time. Annual budgets are discrete activities, while IT activities and spending are continuous. It’s no wonder that arranging funding for things like reducing technical debt is hard when the CIO surprises the CFO with the need annually rather than conducting a continuous, ongoing discussion about risks the company is carrying and how to remediate them.

Annual audits represent discrete cycles for compliance tracking, where we could be practicing continuous compliance, policy enforcement, and auditing. Discrete, usually annual, performance reviews are artificial and sometimes miss opportunities to adjust performance and relationships along the way. Resetting employee salaries once a year means that a competitor can steal the employee with a higher offer during the year. Vetting the security of an IT system periodically allows vulnerabilities to persist until the next opportunity to find them.

Many of the things we do continuously now only became continuous because new enabling technologies were introduced or became widely accepted, like the cloud, automated testing and deployment, and fancy analytics. But many of the technologies needed to make today’s discrete processes continuous already exist; others will come into being once it becomes clear they are needed.

Innovation is an interesting case. Companies often think of innovation in terms of the periodic introduction of new products, product categories, or business models on a grand scale; the sudden eruption of generative AI today may be an example. Because they view innovation as a periodic and dramatic event, they manage and structure innovation that way. They create innovation teams, establish innovation steering committees, search for big ideas, and maybe hire consultants to figure out why they aren’t disrupting their industries.

This type of grand innovation is powerful and is usually the basis for creating a new startup; it’s the dream of venture capitalists to spot the grand innovations and fund them. But grand innovation is also risky and requires gatekeeping and significant risk management efforts (venture capital itself is a powerful gatekeeping function).

The digital world offers the alternative possibility of continuous innovation and indeed encourages it. In this view, employees are culturally focused on continuously innovating to solve problems for their customers and to improve their company’s execution abilities. Innovation doesn’t only mean big, world-changing new product releases but also encompasses an attitude toward daily work—a constant application of new thinking to everyone’s tasks. Innovation is every employee and manager’s job and doesn’t require special permission.

Continuous innovation requires a very different type of management. Its risk needs to become institutionally low—technically, through the creation of “sandbox” environments for experimentation, and socially, by encouraging and expecting that sort of experimentation. Controls that restrict large, risky innovations must be relaxed or streamlined (often through automation) to allow low-risk, incremental innovations to flourish. And leaders need to express a vision within which employees will continuously innovate, one that includes a view of customers and their needs and the company’s basis of competition. Continuous, rather than discrete, innovation is consistent with the general trend toward continuous activity in the digital world.

It may be helpful for enterprises to think of even generative AI through this lens. In the excitement about its still unknown potential, it’s natural to think about AI’s big, disruptive—and maybe risky—possibilities. But over time, AI will become a tool for continuous innovation as well. Today we think of it as something new or separate from our day-to-day way of serving enterprise needs through technology. But every indication is that AI will become the lens through which we see the use of technology in business, as the cloud has become today.

In that sense, we need to manage differently to encourage continuous innovation with AI. We need to reduce its risk through automated controls and attention to ethical considerations. We need to make it easily accessible to employees in doing their jobs—what we at AWS call democratizing AI. We need to establish repeatable processes for reliably, securely, and resiliently deploying AI capabilities. And we need to constantly reference customer needs and opportunities for improving the business to stimulate the flow of innovative ideas. AWS’s AI toolset, including Amazon SageMaker, Amazon Bedrock, and our other AI and non-AI services, will facilitate incorporating AI into continuous innovation.

This is not to say that dramatic, disruptive innovation is unimportant. But traditional companies for which the disruptive possibilities of the digital age may seem frightening, should note that there are also incremental, lower-risk approaches to innovation. And those approaches follow a customer-centric view of business, where innovation is the daily activity of inventing ways to serve customers.

Enabled by technology, business practices are changing. One way to characterize this change is to see it as a movement from discrete to continuous activity. Digital transformation is centered around transitioning to this new approach and making the cultural, organizational, and technical changes necessary to support it. If you are facing challenges in the transformation, you might want to pause and ask yourself whether there are things you do that are still anchored in discrete ways of thinking.