.svg)
.jpg)
Across every industry, leaders are pouring time and budget into modernization—new data platforms, warehouse rebuilds, cloud migrations, and analytics tools. Yet when they step back, the story sounds the same:
"We’ve been modernizing for years, but it still feels like our data can’t keep up with the business—or with AI."
In our recent webinar, Launch Consulting’s Russ Whitman sat down with Dan McKinney (CTO & CIO at Magna Legal Services), Jonathan Gardner (Technology Lead, Launch), and Q Suliman (Partner Technology Strategist at Microsoft) to unpack why so many organizations are modernizing without moving forward—and how to escape that cycle.
The stakes are high. As we highlight in the downloadable infographic that accompanies the webinar, 88% of AI pilots fail to reach production and 95% of gen AI initiatives never deliver meaningful returns. Not because leaders lack ambition, but because the foundation underneath AI is fragmented, brittle, and constantly being rebuilt.
We call this pattern the Modernization Treadmill—and this article breaks down what it is, why it’s so hard to step off, and how the Leapfrog approach can help you turn data chaos into AI opportunity.
Here’s how to build an enterprise data strategy that works.
What Makes the Modernization Treadmill So Hard to Escape?
If you recognize yourself in this picture, you’re not alone:
Jonathan described it as a never-ending cycle: “It’s like you’re constantly on a treadmill of upgrades just to keep accessing the data the way the business needs it.” The frustrating part? As soon as you finish a major modernization phase, the landscape has already shifted. AI advances. Business expectations jump. New data sources appear. And suddenly, you’re behind again.
But the treadmill is more than technical work—it’s systemic inertia created by:
1. Architectural Fragmentation That Compounds Over Time
Most organizations are juggling decades of architectural decisions layered on top of each other: legacy systems, half‑completed migrations, stopgap integrations, departmental tools, and recently adopted cloud services.
Every new business initiative introduces more systems, more data, and more complexity. Instead of simplifying the estate, modernization often adds yet another layer to manage.
2. Business Pressure That Outpaces Technical Capacity
As Dan noted, data teams are being asked to serve departments with wildly different needs—operations, finance, legal, sales, product, compliance. Each requires new datasets, new pipelines, new dashboards.
Modernization rarely gives teams time to stabilize. They’re building the plane, flying it, and redesigning the engines—simultaneously.
3. Exploding Unstructured Data That Old Architectures Can’t Handle
The growth curve isn’t just steep—it’s overwhelming. Magna manages transcripts, PDFs, audio, video, and case data that multiply daily. Historically, BI systems were built around structured data. Now AI gives value to everything, which means everything must be processed, classified, governed, and made retrievable.
Jonathan emphasized that this turns linear data growth into logarithmic demand. Traditional rebuild-and-refactor models simply can’t keep up.
4. Modernization Efforts That Reinforce, Rather Than Solve, the Problem
Teams often rebuild systems “the right way” only to discover:
So, the cycle restarts.
This is why modernization efforts often generate activity without generating progress. Teams celebrate pipeline rewrites and warehouse consolidation—but the system underneath remains fragile.
Jonathan described it perfectly: “It’s like being in a permanent state of renovation. By the time you finish one room, the rest of the house is already outdated.”
Why Traditional Modernization Fails in an AI-First Era
AI raises the bar for what data must do:
Traditional modernization assumes you have time to rebuild everything first. But the AI era won’t wait.
Traditional vs. Leapfrog: Why Leaders Are Switching Approaches
The infographic captures this shift simply but powerfully:
Modernization used to be infrastructure-first. AI requires intelligence-first.
This isn’t about upgrading systems for the sake of upgrades. Leapfrogging is about escaping the cycle entirely—moving from rebuilding to rethinking. It allows organizations to bypass years of technical debt, align their data strategy with AI, and shift from reactive maintenance to proactive intelligence.
In short: stop laying more “phone lines” in a 5G world. Start building for the world you’re operating in now—and the one AI is rapidly shaping.
more “phone lines” in a 5G world. Start building for the world you’re operating in now—and the one AI is rapidly shaping.
What Does AI Change About Your Data Strategy?
The panel started with a simple premise: there is no good AI without good data. But AI doesn’t just place another demand on your data architecture—it fundamentally changes what that architecture needs to do.
Dan shared Magna’s reality: a regulated legal services business managing massive volumes of structured and unstructured data—PDFs, transcripts, video, audio, and more. Historically, teams built analytics around structured data from source systems. Now, AI opens up unstructured content as well, which means:
At the same time, as Jonathan pointed out, data growth has gone from linear to exponential to what feels almost logarithmic. Every new product, every new use case, every new AI experiment generates even more data. The old approach—just building more warehouses and more pipelines—simply doesn’t scale.
That’s where the Leapfrog model comes in.
How Does the Leapfrog Approach Solve Data Chaos Faster?
Instead of asking, “How do we rebuild everything?” the Leapfrog approach asks:
“How do we build an AI-ready foundation as quickly as possible—without tearing the house down?”
The model breaks into three major “leaps,” all of which show up visually in the webinar’s infographic:
Together, they give you a way to step off the treadmill and move toward AI-readiness in months—not years.
1. How Does a Unified Data Fabric Help You Finally See Your Business Clearly?
The first leap focuses on connection, not reconstruction.
A unified data fabric—like Microsoft Fabric built on OneLake—sits across your existing systems and creates:
Q described OneLake as “OneDrive for data”: a single logical lake that gives you one copy of data and many ways to access it. Instead of copying the same data into multiple silos for every analytics and AI use case, you centralize it once and then build from there.
For organizations stuck in “pipeline purgatory,” this is a huge mindset shift. You don’t have to move everything before you see value. You can:
As Russ put it during the session: “You can start delivering intelligence now—not after a three-year rebuild.”
2. What Happens When You Replace Manual ETL With AI-Ready Data Products?
The second leap transforms how you engineer data.
In traditional environments, ETL is where projects go to die. Pipelines multiply, hand-coded transformations become fragile, and every new use case means another round of rebuilding. That’s why the infographic highlights AI-ready data products as a core Leapfrog move.
With AI-ready data products, you shift from bespoke pipelines to:
For Magna, that means going beyond simply storing data to actually making it productized and contextual. Dan talked about using data to power:
Those scenarios only work if the data products underneath them are consistent, fresh, and governed. Once they are, AI agents and applications can finally operate with confidence.
3. How Does Self-Service Intelligence Break Your Dependence on IT?
The third leap is where modernization becomes visible to the business.
Self-service intelligence tools—especially Power BI combined with Copilot—let people interact with data conversationally:
For Dan’s world, this means that not only the data team, but schedulers, operations leaders, and client-facing teams can get answers on their own. For Q and the Microsoft side, Fabric is intentionally designed so that less technical users can also tap into AI-powered experiences, not just data engineers and architects.
Jonathan summed it up nicely: “When insight becomes conversational, intelligence becomes scalable.”
You’re no longer modernizing just to produce better dashboards. You’re modernizing to give every decision-maker a smarter, AI-augmented copilot.
Where Should You Start If You’re Already Mid-Modernization?
A big question from webinar attendees was: “What if we’re already deep into a modernization program—do we have to start over?”
Short answer: No. But you probably do need to rethink the path forward.
Dan’s guidance was to start with low-hanging fruit and use it to build “muscle memory” for AI:
At Launch, Jonathan and Russ talked about how they use AI itself to accelerate this discovery process. Instead of sending a squad of analysts to manually reverse-engineer your pipelines and codebase, Launch uses AI-powered diagnostics and agents to:
What used to take several weeks of assessment can often be compressed into about a week—and in many cases, teams start making tangible changes during that same time window.
The goal isn’t to add another pilot to the pile. It’s to build muscle, not just pilots—and to create a repeatable way to spot and execute high-value AI and data opportunities.
How Are Leaders Planning for 2026 in an AI-First World?
Looking ahead, Dan shared a challenge many leaders are wrestling with: how to budget for a future where AI is both a cost line and a force multiplier.
Instead of just thinking about tokens or model pricing, Magna is evaluating:
Internally, Magna now views AI as a force multiplier across roles—from IT and security to investigators and trial consultants. The question isn’t, “Can AI replace this person?” but rather, “How can AI help this person support more clients, more work, and more complexity than ever before?”
That mindset aligns with how Launch approaches AI with clients: keep humans in the loop to direct, verify, and apply AI—while using AI to extend their reach.
Why Does the Leapfrog Mindset Matter for AI’s Future?
The webinar—and the infographic—both land on the same conclusion:
You don’t have to modernize the old way.
Organizations that adopt the Leapfrog mindset experience what the infographic calls “The Payoff”:
Instead of trying to perfect an architecture that will be outdated the moment you finish, you:
This isn’t about starting over. It’s about starting smarter.
FAQs
Why do so many AI pilots fail?
Most fail because the underlying data foundation is fragmented, poorly governed, or hard to access. AI can’t succeed on top of siloed, brittle data structures.
Does Leapfrog modernization replace traditional cloud migration?
No. It accelerates and focuses it. Leapfrog lets you unify and activate data earlier, so AI and analytics can start driving value while deeper migrations continue.
How quickly can an organization become AI-ready?
With the right patterns and platforms, organizations can achieve meaningful AI readiness in as little as 90 days, even if broader modernization efforts are still underway.
Is Leapfrog only for “greenfield” environments?
Not at all. In fact, it’s designed for complex, real-world environments with a mix of legacy systems, cloud platforms, and SaaS tools. The goal is to connect, not restart.
What kind of use cases are best to start with?
Look for processes that are data-heavy, repetitive, and business-critical—like scheduling, document review, or customer support. These often deliver quick wins and clear ROI.
Modernize Once. Leap Forever.
Modernization doesn’t have to be a prolonged, expensive, and frustrating cycle. With the Leapfrog approach, you can step off the treadmill, build an AI-ready foundation, and start turning data chaos into competitive advantage.
Instead of rebuilding everything for the third or fourth time, you can:
You don’t need to wait for the “perfect” architecture to start.
You can begin your leap today—and modernize in a way that keeps pace with AI, instead of endlessly chasing it.
