Good Infrastructure Is Boring Until It Isn't
Danielle Diliberti on shrinking the stack, compounding data errors, and what the gap between AI leaders and laggards actually looks like up close.
Danielle Diliberti has watched two transformation waves reshape enterprise technology from an unusual vantage point. As a board member of the MACH Alliance and CEO of Sommsation—a lean, fast-moving organization she co-founded and runs with a lean team and an AI-augmented operating model—she's been both an advocate for composable architecture and a practitioner living its consequences in real time.
Her read on where enterprises stand right now with AI is characteristically direct: the organizations that did the foundation work are pulling ahead, and the distance between them and those that didn't is growing faster than most people expected.
Speed Has Changed the Calculus
The parallel Diliberti draws between the composable movement and the current AI moment is instructive, and so is where she says it breaks down.
"What's different about AI from every prior transformation wave is that the technology doesn't necessarily wait for organizational buy-in,” she says. “In the composable era, humans still controlled the pace. That's no longer true, and most operating models aren't built for this speed.”
During the composable era, large-scale transformation still depended heavily on business decision-makers: budget cycles, adoption timelines, organizational buy-in. The pace was controlled, if sometimes frustratingly so. But that dynamic has inverted. The technology now moves faster than many organizations can react to, and the implications cut both ways.
The upside is real—building and discarding experiments costs less in time, money, and energy than it used to—but the corollary is harder to manage. QA at this speed becomes genuinely overwhelming, and unwinding things that don't work requires visibility that many enterprises simply don't have. The token economy is forcing IT and finance into unfamiliar proximity, and the expanded capability of engineers is beginning to outpace the business structures designed to govern them.
"The speed is really impressive," Diliberti says. "Figuring out how to manage that is the challenge."
What Good Foundations Actually Buy You
"The inefficiencies are visible and able to be untangled and rectified easily, but only if the framework was built that way to begin with."
"Our leadership team was careful and smart about infrastructure design, so there was a framework that relied on quality in, quality out," she says of how Sommsation has been able to move faster than most orgs. "It meant there was a leg up years ago that others maybe didn't have."
The practical result is their team has visibility into where inefficiencies live, and the ability to untangle them without disturbing everything else. For organizations whose foundations were more hastily assembled, that visibility simply isn't there, and AI compounds the problem rather than compensating for it.
The data architecture point is where Diliberti gets most specific. She says that organizations that haven't paid enough attention to data foundation work are missing important context. In agentic systems, where agents are communicating across multiple layers, a single incorrect data point connected to other data points becomes a proliferation of error that adds up quickly. With humans in the loop, you catch it and move on, but without the context and domain experience to know what to look for, the gap widens silently because businesses don’t even know they have a potential problem.
"People won't be aware of it," she says. "That gap in context is going to be the kicker on a lot of that stuff."
The Sommsation Model: Smaller Stack, Leaner Team, Faster Everything
Diliberti makes the case for strong foundations practically as much as theoretically. At Sommsation, the composable architecture built into the organization from the start has not only supported AI adoption, it's made a deliberate, considered response to it possible.
"We've started simplifying the operating model," she says. "Building a modern architecture from the onset gave us that optionality. AI is evolving where we want to make operational changes that make sense, and it's not just in technology decisions.”
The point isn't that composable stacks should shrink, Diliberti notes. It’s that composable architecture gives organizations the visibility and flexibility to make those calls on their own terms, adding or consolidating capability as business needs evolve rather than being locked into decisions made under different conditions. That way you know what's safe to change without pulling a thread that derails everything else.
For Sommsation, that means streamlining certain functions while keeping the underlying composable foundation. It’s what makes the whole operating model coherent and adaptable, after all.
"When we first launched, we knew that the current industry infrastructure wasn't going to keep pace with us as we grew, and we knew we'd carry technical debt as a result," Diliberti says. "We had to build bandaids by design." Those compromises were made with clarity up front, and were what made the composable infrastructure so valuable to their team.
That kind of architectural self-awareness—knowing what you're building, why, and what you'll eventually need to revisit—is exactly what composable foundations make possible. Skipping that work means you don't have the same visibility into what to change, or the same confidence that changing one thing won't break everything adjacent to it.
The Honest Assessment
What Diliberti is describing is a realistic accounting of where the market stands. Some organizations are well-positioned and moving fast. Others are discovering that the shortcuts they took on data architecture, vendor governance, or organizational design are now material liabilities.
The window to close that gap isn't closed, but it is narrowing, and the speed of the technology isn't waiting for anyone to catch up.
Leigh Bryant
Editorial Director, Composable.com
Leigh Bryant is a seasoned content and brand strategist with over a decade of experience in digital storytelling. Starting in retail before shifting to the technology space, she has spent the past ten years crafting compelling narratives as a writer, editor, and strategist.