Building an AI-Enabled SDLC: Insights from 3Pillar and Forrester

As AI reshapes how software gets built, many organizations are asking the same question: where do we start—and how do we make it real?

That was the focus of a recent joint webinar hosted by 3Pillar and Forrester, titled AI & the SDLC: Beyond the Hype to a Practical Roadmap. In a candid discussion, Lance Mohring and Scott Young, Field CTOs at 3Pillar, joined special guest Devin Dickerson, Principal Analyst at Forrester, to explore what true AI-enabled velocity looks like – and how enterprises can prepare for it.

The conversation surfaced a clear message: success in the AI-driven SDLC isn’t about chasing tools. It’s about understanding maturity, orchestrating change, and building the right foundation for long-term impact.

To ground the discussion, 3Pillar introduced its AI-enabled SDLC maturity model – a framework designed to help technology and product leaders assess where they stand today and define the next step forward.

The model maps the evolution from today’s fragmented experimentation to tomorrow’s orchestrated, AI-native development environments:

• Level 0 – Classic SDLC: Traditional agile, kanban, and waterfall practices. Stable and familiar, supported by a mature tool set, but largely manual.
  
• Level 1 – AI-Assisted: Individual practitioners begin experimenting with tools like code assistants, design generators, or test automation. Progress is real but inconsistent.
  
• Level 2 – AI-Optimized: The organization formalizes adoption; governance, tool selection, and integration create measurable efficiency gains across the lifecycle.
  
• Level 3 – Agent-Augmented: Planning agents begin managing end-to-end workflows, coordinating across roles and tools. Humans move “over the loop,” providing direction and validation. Scrum is supplanted as the dominant work management method.
  
• Level 4 – Agent-Native: Teams operate in fully AI-orchestrated environments where human expertise focuses on strategy, design, and value creation. The process is optimized for AI, with the humans focusing on inputs and outputs

“We built this to give leaders a way to visualize progress – to see not just what’s possible, but what’s practical today.” – Scott Young

While the framework spans five levels, most organizations remain early in the journey. Dickerson observed that maturity is uneven by nature.

“You might have a few developers running sophisticated agent workflows while others on the same team don’t see the value yet. AI maturity isn’t linear—it’s fragmented.” – Devin Dickerson

In other words, enterprise adoption today looks more like a patchwork than a transformation. Designers may experiment with AI for ideation; developers may use copilots for small code snippets. But without alignment, these efforts plateau quickly.

Young illustrated the point through a real example: after using an LLM to prototype a React app, he handed it off to a designer – who had no idea how to integrate it into Figma. “We’re getting velocity at the individual level,” he noted, “but we’re not scaling collaboration.”

This is the defining marker of Level 2: impressive local gains that don’t yet translate into systemic value.

To explain this dynamic, Young used a powerful metaphor. In the early 1900s, when Henry Ford asked people what they wanted, they said “a faster horse.” What they really needed was a car.

Many organizations are in a similar moment today – seeking incremental speed in legacy processes rather than reimagining the system itself. They’re improving story writing, test automation, or code generation, but still within the same SDLC architecture.

“The real shift happens when you stop thinking about tools and start thinking about orchestration.” – Lance Mohring

When orchestration replaces isolated experimentation, the returns multiply. Moving from AI-assisted to agent-augmented workflows can unlock 10× efficiency gains in focused use cases.

At this stage, humans shift from “in the loop” to “over the loop.” Instead of manually prompting each task, they oversee a network of agents performing coordinated roles – architect, tester, product manager, developer – each operating semi-autonomously.

This new structure changes more than output speed. It collapses the traditional seams of the SDLC, accelerates feedback loops, and forces organizations to rethink concepts like backlogs, sprints, and even technical debt.

“Technical debt may become less relevant. The AI doesn’t care that your code isn’t elegant—it just makes the change everywhere instantly.” – Scott Young

But acceleration without governance can magnify risk. Trust, quality, and validation processes must evolve just as quickly. “Did we build the right thing?” remains the most important question.

In the next phase of maturity, the SDLC becomes less about humans instructing narrow task agents and more about planning agents coordinating entire workflows.

A task agent might add a new field to a form; an planning agent could translate a product requirement document into a full sprint backlog, direct subordinate agents to execute each story, and verify completion – all under human supervision.

“We’re moving from delegating tasks to delegating responsibility—while humans retain accountability.” – Devin Dickerson

This evolution redefines roles. Product managers, architects, and designers become system conductors rather than task executors. The orchestration layer becomes the new frontier of competitive advantage.

The speakers closed with four practical recommendations for leaders beginning this journey:

1. Start now. Experiment with code-generation use cases to build understanding of what AI can – and can’t – do.
   
2. Prioritize proficiency. Invest in context engineering and prompt literacy. The real gains come from knowing how to guide AI effectively.
   
3. Identify high-impact use cases. Focus on choke points – testing, QA, or backlog refinement – where automation drives visible results without high risk.
   
4. Keep people at the center. Balance governance with creativity. Reinvest productivity gains into upskilling your teams and nurturing cross-functional collaboration.

“AI in the SDLC isn’t a single project – it’s an organizational transformation that demands structure, experimentation, and empathy in equal measure.”

The complete conversation between 3Pillar and Forrester offers far more depth – including examples of real-world implementation and predictions for what’s next in AI-native product development.

Watch the on-demand webinar