Value Stream Management (VSM) was all the rage a few years ago, but now there’s a new contender for measuring how much value software is driving: Software Engineering Intelligence. 

“While Value Stream Management continued to lose steam in 2024, we also saw the fast emergence of Software Engineering Intelligence (SEI) to take its place,” Derek Holt, CEO of Digital.ai, told SD Times while making predictions for 2025. 

Holt was interviewed in the most recent episode of our podcast to talk more about this trend, explaining that SEI is the application of both data driven insights and predictive AI to measure and improve not only the process of delivering software, but also the outcomes of that software development.

“This includes both rear view metrics of how things have trended in the past, but also looks at predictive insights to be able to identify areas for improvement, risk avoidance, etc,” he said. “There’s a reason why the rear view mirror is smaller than the windshield in a car, and we think it is that balance of both views that are going to help organizations just get better and better at predictably, building and delivering great software.”

There are a couple of different reasons why Holt believes SEI will gain more traction in the year ahead. 

First, organizations are continuing to look at how to mature the way they build and deliver software. 

Second, there is pressure on companies to adopt AI in their software development process, but they don’t necessarily have the data to justify whether or not AI has resulted or will result in improvements. 

Third, we now have the technology stack needed to gather structured and unstructured data from third-party tools and provide it in a dashboard and perform predictive insights on that data.

Looking back to VSM, it really was designed to help answer two questions:

  1. How good are we at building software?
  2. How good is the software we build right now?

“I know those are a lot of the same words, but they’re wildly different things,” Holt explained. “I could be really great at building, let’s say, a new smartphone, but if the smartphone isn’t wanted by anybody, it doesn’t really matter that I was good at building it. And at the same time, I could have the best idea for the next generation smartphone, but if I don’t have a way to get the parts, manufacturing, and distribution, that idea is not worth a lot either.”

Holt believes that SEI is a part of the overall vision for VSM. It’s not a complete departure from VSM, but rather a recognition of that first question. 

He explained that development teams often have an easier time understanding how good they are at building software, and a harder time understanding the value they derive from that software, as that can vary based on business model, business strategy, market conditions, etc.

Holt also emphasized that it’s not an “either or” when it comes to VSM vs SEI. “I think it’s a subset, and hopefully we continue to build towards the original vision of value stream management, just in a kind of a step function over time.”

For teams looking to implement SEI, Holt says there are a few prerequisites. First, whatever vendors they’re working with need to have ways to pull data from those systems, as no singular data source can provide that 360 degree view that is needed. 

Another important foundation is the ability to scale. “These things are easy to demo,” Holt said. “They’re harder to scale across an enterprise. So I would really encourage folks, as they’re doing evaluations, to make sure that they’re partnering with folks that have done this at scale. And that’s not just scale of data, it’s also scale of users, scale of continuing to mature … getting value right away, but also then expanding the purview in terms of the types of intelligence that you’re getting.” 

He also went on to say that it’s important to have that 360 degree view so that you can ensure you’re not only able to look back, but also predict future outcomes. 

His final piece of advice is to not overanalyze the options in a way that prevents you from getting started. “It’s very, very easy to look at this space and to spend a year debating and having a bit of analysis paralysis … But the very first view you get, the very first aggregation of multiple data sources that gives you insights into your business, even just one is better than where you were before. And so this is a journey, not a destination, and a key piece there is taking that first step. So I would just encourage folks to get started instead of waiting for perfection.”