How Agile development principles can help with scaling AI deployment
In this podcast, Andrew Sales, chief methodologist at Scaled Agile, says lean and Agile methods can help companies take full advantage of AI's transformative impact on work.
Organizations are coming to realize that AI represents a major opportunity to not just do existing work faster, but to change what kind of work is done and who does it. It requires a comprehensive plan and commitment to redesign work.
Well-tested models for this kind of overhaul of tasks and workflows exist in the Six Sigma, lean management and Agile development methodologies companies have long used to build anything from chairs to smartphone apps. Now purveyors of those tools, frameworks and methodologies -- and the training and certification needed to apply them -- are turning their attention to the challenges of deploying AI on an enterprise scale.
In this episode of Enterprise Apps Unpacked, Andrew Sales, chief methodologist at Scaled Agile, explains his company's new AI-Native SAFe framework and how it can help redesign product decisions, including related governance, workflows, and team collaboration, before AI is added to the mix.
Andrew Sales
Lean and Agile AI
Scaled Agile is influential in the Agile software development movement that formed at the turn of the century. The company claims to have trained more than two million people in its Scaled Agile Framework (SAFe) approach for scaling lean and Agile practices. Major government agencies like NASA and roughly two-thirds of the Fortune 500, including Boeing, Cisco, CVS Health and FedEx, are among the 20,000 organizations around the world said to use SAFe.
While most of Scaled Agile's customers have used SAFe for internal software development and other digital systems, Sales said the method can be applied to any kind of product. "When we think back to lean and Agile, it was predominantly in the space of technology," Sales said. "What's different here is we're seeing this as a whole-organization transformation. We've started to see people from all departments come on our training: legal, marketing, finance -- anyone who has a workflow."
The new framework, which is built on but doesn't replace the core SAFe framework, is primarily intended to help companies deploy AI applications more thoughtfully and on an enterprise-wide scale, but it also helps them understand and take advantage of the ways AI changes existing development processes.
Sales said lean and Agile focused on efficiencies, time to market, and accelerating delivery of new features to customers, while AI is changing the goal of product development from outputs to outcomes. AI has also removed bottlenecks and capacity constraints across entire organizations.
"We're seeing smaller teams, AI-augmented teams of three to four people building products in time frames that used to take many, many teams, many weeks to build," he said. And that, in turn, creates challenges in managing the increased output. The answer is to not worry about managing output and instead view overproduction as integral to innovation. That means workflows now must include a process for discarding work that doesn't serve the stated outcomes.
Other topics discussed in the podcast include the following:
- The key elements of AI-Native SAFe.
- Evidence that the method works.
- Research that identified common mistakes in AI deployment.
- Best practices.
David Essex is an industry editor who creates in-depth content on enterprise applications, emerging technology and market trends for several Informa TechTarget websites.
