Enterprise transformations are having a moment – and this time, the opportunity is real. ERP upgrades, cloud migrations, and AI-enabled operating models are converging in a way that gives leaders a genuine chance to rethink how their organizations work. At the same time, cost pressure, talent constraints, and technology complexity are all increasing. The upside is enormous, but the margin for error has never been smaller.

This Q&A captures a wide-ranging conversation between acceligence CEO, Justin Greis, and Mike Casey, executive advisor. What follows is shaped by more than three decades of leading, rescuing, and in some cases stopping large-scale enterprise transformations. Mike has  seen the same patterns repeat across industries, technologies, and generations of tooling, and the examples he shared are meant to be practical, not theoretical.

If there’s a unifying belief behind everything that follows, it’s this: when enterprise transformations are approached with clarity, discipline, and honesty, they can deliver far more than new systems. They can unlock business value, modernize how work gets done, and develop the next generation of leaders.

What are the top mistakes and missed opportunities when implementing enterprise applications?

“Most transformations don’t fail because the problems are unknown. They fail because the known problems get ignored.”

Justin: You’ve seen dozens of enterprise transformations up close. When you walk into a program, what are the common mistakes and missed opportunities that show up again and again?

Mike: I’ll start with the uncomfortable truth. Most enterprise transformations don’t fail because of exotic technical issues. They fail because organizations already know what matters and then choose not to act on it once the program gets moving.

The biggest issue, without question, is leadership and sponsorship paired with real business involvement. Executive support can’t be ceremonial, and business participation can’t be delegated away. If the business thinks this is an IT project, you’re already in trouble. I’ve seen technically sound programs deliver very little value simply because the business never truly owned the outcome or stayed engaged past kickoff.

Data is the next fault line. ERP systems have always lived and breathed on data, but AI has raised the stakes dramatically. Weak master data, fragmented governance, and unclear ownership don’t just undermine ERP performance; they actively block AI adoption. I’ve worked with organizations that struggled with data during their original ERP implementations and are now facing the exact same issues as they try to layer AI on top. The opportunity today is that AI itself can help remediate and enrich data faster than anything we’ve had before, but only if leaders are willing to confront the problem directly.

Methodology and resourcing are another common breakdown. Programs get into trouble when delivery approaches and talent models fall out of balance. Strong people can sometimes compensate for weak methodology, and strong methodology can sometimes compensate for uneven teams. But when both are misaligned, scope, quality, and budget erode quickly. Agile and scaled agile don’t save you if they’re treated as ceremonies instead of discipline.

Another mistake I see constantly is confusion about transformation intent. Organizations blur the lines between big T transformation, little T transformation, and technology-only transformation. Problems show up when leadership believes they’re pursuing one thing and execution quietly drifts into another. This is why I’m a big believer in program charters that are actually used, revisited, and refreshed. A charter isn’t paperwork; it’s the “North Star” for thousands of daily decisions.

Two modern pitfalls deserve special attention. One is architecture by vendor, where the enterprise blueprint quietly turns into a product catalog shaped by supplier incentives instead of business needs. The other is tech debt transfer – lifting legacy complexity into new platforms without rethinking how systems, processes, and data should work in a modern environment. Done poorly, you don’t reduce cost or complexity. You just lock it in and start paying for it on a subscription model.

How do you infuse AI into your ERP stack in a way that actually creates value?

“AI isn’t hard because it’s powerful. It’s hard because there’s too much of it, moving too fast.”

Justin: AI is the big elephant in the room right now. How do organizations actually start infusing AI into their ERP and enterprise application stacks without getting lost in the hype?

Mike: What I see most often isn’t a lack of ideas; it’s paralysis. There’s simply too much happening, too fast, and too many choices that all look different. If you don’t structure how you approach AI, it’s very easy to wait, dabble, or chase the wrong thing.

I always start with awareness before tools. Organizations need a disciplined way to scan the AI ecosystem and understand what’s real and usable today, not what’s promised on a roadmap. That means knowing where models are production-ready, how they integrate with your existing platforms, and what they actually cost to operate. I often compare this to streaming services: you sign up for one thing, and over time you’re paying more and more for features you didn’t plan on. AI and cloud platforms can behave the same way if you’re not intentional.

From there, I think about AI value in three layers. The first layer is embedded AI – capabilities already built into enterprise applications. These deliver modest efficiency gains with minimal disruption. They’re often easy wins and a good place to start, but they’re rarely transformative on their own.

The second layer is agentic and workflow-oriented AI. This is where you start automating end-to-end activities, reducing manual handoffs, and letting systems run continuously while humans focus on exceptions and tuning. This is where double-digit productivity improvements become realistic, provided your data, processes, and governance are ready.

The third layer is reinvention. This is where you step back and fundamentally rethink how work gets done by combining business insight with AI capabilities. These efforts are fewer, more expensive, and harder to execute, but this is where breakthrough value lives. Reinvention shouldn’t happen everywhere at once. Selectivity is a feature, not a flaw.

Across all three layers, the principle is the same. AI works when it’s applied deliberately, anchored in business outcomes, and supported by architecture and operating models that can evolve as the technology continues to change.

Do you need clean data before using AI, or can AI help fix the data problem?

“ERP systems are data-heavy, but they’re often information-poor.”

Justin: Do organizations really need pristine data before they can use AI, or can AI actually help solve the data problem itself?

Mike: This is one of the most common misconceptions I run into. Historically, ERP implementations optimized for minimal viable data – just enough structure and quality to process transactions reliably. That discipline made systems stable, but it also starved analytics and machine learning of the richness they need. As an industry, we spent 20 years making ERP data lean. To get to real insight, AI needs data to be expansive.

Advanced AI models need expansive, contextual data, and much of that data doesn’t live inside the ERP. Weather patterns, market signals, supplier behavior, and customer dynamics all matter, but they typically exist outside core systems. The good news is that modern cloud platforms and AI-enabled data tools dramatically reduce the cost and effort required to expand, cleanse, and harmonize data.

AI copilots can now assist with profiling, enrichment, and quality improvement, allowing organizations to move forward without waiting for some mythical state of data perfection. Rather than insisting on a single system of record, I see organizations moving toward a model where data is physically distributed but logically unified through semantic layers.

In that model, governance and meaning matter more than physical co-location. AI becomes both a consumer of the data fabric and a catalyst for improving it.

What battle scars stand out from the world’s largest transformations?

“If you ever have to explain your ERP program to Wall Street, something went very wrong.”

Justin: You’ve said some of your proudest moments were stopping programs that needed to be stopped. What battle scars stand out most?

Mike: The failures that make headlines are almost never sudden. They’re the result of fundamentals being ignored for too long.

Business readiness is a huge tipping point. I’ve seen years of solid technical work undone because training was incomplete, processes weren’t understood, or legacy behaviors carried into new systems. In one situation, warehouse teams continued using manual processes on top of a new system of record, which led to inventory discrepancies that eventually had to be explained publicly. You can’t run old processes on new systems and expect new outcomes.

Methodology breakdowns are another recurring pattern. When you have multiple partners, inconsistent delivery approaches, or weak integration discipline, risk quietly accumulates. Teams optimize locally, defects shift downstream, testing gets compressed, and quality erodes under pressure.

Some of my proudest moments involved stopping programs altogether. In several cases, fit-for-purpose analysis made it clear that heavily customized package solutions were fundamentally misaligned with the business. Continuing would have locked the organization into a high-cost, low-value platform for years. Sometimes the best transformation decision is deciding not to transform the wrong way. Ending those efforts early was painful, but it preserved long-term value.

How do you get the most value from consultants and trusted advisors?

“Consultants aren’t the problem. Confusion about who owns what is.”

Justin: What advice would you give leaders about working with consultants and actually getting value from them?

Mike: Consultants are tools. They’re neither heroes nor villains. Their value depends entirely on clarity of ownership, accountability, and capability.

Everything starts with self-awareness. Leaders need to know which capabilities belong in-house and which should be sourced externally. Confusion at those boundaries almost guarantees disappointment. If the client can’t manage the partner, the partner will end up managing the client.

I also challenge the assumption that one large integrator should own everything. As architectures become more modular and programs shorten, there’s a real opportunity to use different partners at different phases, as long as the client retains stewardship of the overall design and delivery model. It is, afterall, a new “home” built for the client’s future residence; it should be treated as an asset throughout the development process, with the right tradesmen, artisans, and decorators assigned at the right time. Some partners provide their own senior executive quality assessors, but wise clients often involve independent advisors to define the issues in quantified and actionable dimensions.

Procurement plays a critical role here. Too often, the process optimizes for lowest cost instead of fit-for-purpose capability, which almost guarantees downstream issues. Procurement should qualify partners first and negotiate price second.

The best engagements treat charters, contracts, and change control as living instruments, not documents that disappear after signing.

What should CEOs know before, during, and after an ERP transformation?

“If transformation sits too low on the priority list, it won’t fail loudly. It will fail quietly and expensively.”

Justin: If you put on your CEO hat, what should leaders know before, during, and after a major ERP or enterprise transformation?

Mike: First, be brutally clear on intent. Is this a fundamental business transformation, a targeted modernization, or a technology remediation? Each demands different levels of senior talent, risk tolerance, and commitment. You don’t staff the biggest change in your company with your second-best people.

During execution, leaders need to assess the organization’s true capacity and readiness for change. Competing initiatives, external disruptions, and unresolved uncertainty all drain momentum. Treating the program like a temporary department, with clear incentives and accountability, helps sustain focus.

After go-live, the people agenda becomes decisive. These programs are rare opportunities to develop future leaders by immersing them in cross-functional, high-stakes work. These programs can be career killers or career accelerators. The difference is intentionality.

If you stay engaged at the right altitude – not micromanaging, but never disengaging – you dramatically improve the odds that the transformation delivers real, lasting value.

A closing thought

“The goal isn’t to avoid mistakes entirely. It’s to make new ones on purpose, with eyes wide open.”

Mike: After decades in this space, the pattern is clear to me. Organizations don’t struggle because transformation is impossible. They struggle because they underestimate the depth of change required and overestimate how forgiving reality will be.

When these programs are grounded in a clear charter, staffed with the right talent, supported by honest partnerships, and treated as business change rather than technology delivery, they can unlock extraordinary value. They modernize operations, develop leaders, and fundamentally improve how a company works.

My hope is that these insights help leaders shorten the distance between intent and impact and get the most out of the enterprise technology transformations they’re already committed to.