media-blend
text-black

Digital Boardroom

AI vs. ERP: Why you’re asking the wrong question

AI and ERP aren't rivals. Here's why mid-market leaders who invest in both win.

default

#

default

#

primary

default

#

secondary

Every business leader we talk to right now is wrestling with some version of the same question: Should we invest in AI, or get our ERP in order first?

It's the wrong question—and the cost of getting it wrong can be years of competitive disadvantage.

We recently sat down with Darwin Deano, President and Co-Founder of Stanza and former Distinguished Partner at McKinsey & Company, and Joe Boden, Partner at McKinsey & Company, on the Future of ERP podcast. Between them, they've helped hundreds of organizations navigate ERP transformation. Right now, they're watching the same mistake play out: companies betting on AI instead of ERP modernization—and stalling out.

"AI is the how," Darwin told us. "ERP is the what and the why."

That framing changes everything.

Why AI without ERP fails fast

Here's the uncomfortable truth: AI is only as smart as the data you feed it.

Most enterprise data lives in your ERP. It encodes your business processes, your operating rules, your institutional knowledge. When that foundation is fragmented, outdated, or inconsistent, AI doesn't fix the problem—it amplifies it.

Joe put it plainly: "Garbage in with AI—it's faster garbage out."

MIT's NANDA initiative found that 95% of generative AI pilots deliver no measurable profit-and-loss impact. The issue isn't the AI model. It's that a powerful reasoning system built on a broken foundation produces confident-but-wrong answers.

Signs your ERP may not be AI-ready:

  • Data is siloed or fragmented across multiple systems
  • Business processes haven't been updated in years
  • Manual workarounds compensate for broken process design
  • Heavy custom code has accumulated over multiple implementation cycles
  • AI pilots run on hand-curated extracts, not live integrated data

Darwin shared a real example: a consumer company with 84 employees dedicated solely to managing intercompany exceptions. That headcount exists because the underlying process is broken. No AI agent solves that without fixing the foundation first.

How is AI already changing ERP?

Far from making ERP obsolete, AI is reshaping what ERP looks like—and making transformation faster and more accessible than ever. In a recent McKinsey & Company article, Darwin and colleagues identified five structural shifts underway.
Disruption
What it means in practice
1
ERP architecture evolves toward “headless”
Users interact through agents and natural language, not screens
2
ERP modernization still matters
A modern ERP is the prerequisite—not the alternative—to AI
3
Transformations get faster and cheaper
AI-enabled implementations can be up to 50% faster with up to 60% fewer resources
4
ERP vendors may regain ecosystem control
Native AI in proven domain platforms outperforms bolt-on point solutions
5
Value creation shifts from build to buy
Standardized processes should be consumed from vendors, not rebuilt
Point three deserves attention for mid-market leaders. ERP transformation has historically felt like a multi-year, capital-heavy commitment. The barrier of entry for modernization is dropping—and that changes the calculus for every business still running legacy systems.

What should leaders at growing companies do in the next 12 months?

If you're running a growing company, resources are finite and every investment decision matters. Here's the practical advice Darwin and Joe shared.

To modernize with intention:

  • Pursue the "irreducible core"—the lean ERP your business truly needs, without years of accumulated customization weight
  • Treat ERP as a perpetual responsibility, not a one-time project—it's the DNA of your organization, and it must evolve as you do
  • Use AI to make the modernization itself better—it can accelerate data migration, automated testing, and process redesign from day one

To avoid the build-vs-buy trap:

  • Buy the 80%—let your ERP provider absorb the complexity of embedding AI into standardized, foundational processes
  • Build the 20%—invest in capabilities that genuinely differentiate your business in the market

The trap to avoid is island pilots. Disconnected AI experiments deliver demo results and production disappointment. Deloitte's 2025 research found that 47% of AI investors also invested in ERP—recognizing it as the operational backbone AI needs to function reliably at scale.

There is no greater capability within the enterprise that represents the business context, the data, the processes, than an ERP. It's essentially the DNA of your business.
— Darwin Deano, President and Co-Founder of Stanza and former Distinguished Partner at McKinsey & Company

What is the Future of ERP?

We always close the podcast with this question. Darwin and Joe's answers were precise.

Darwin: "ERP will be lean. It will be the system of context as well as the system of record. And it will be the primary vehicle for unlocking AI value at scale."

Joe: "The ERP of the future will be invisible to the average user—but indispensable. It will shift from a system of record to a system of action: not just logging what happened, but proactively telling participants what to anticipate and what should happen next."

Invisible. Indispensable. A system of action.

That's not the end of ERP. That's ERP at its most powerful.

The leaders who win won’t pick one—they’ll invest in both

The leaders who pull ahead in the AI era won't be the ones who chose AI over ERP. They'll be the ones who invested in both—and understood that a clean, modern ERP is what makes AI actually deliver at scale.

One McKinsey & Company client—an OEM manufacturer—unlocked more than a billion dollars in productivity by approaching AI and ERP as a unified domain transformation, not a series of disconnected pilots. The formula isn't complicated. But it does require getting the foundation right.

Listen to the full conversation on The Future of ERP podcast. And if you're ready to explore what AI-ready cloud ERP looks like for your business, visit sap.com/grow.

FAQ

Should I invest in AI or modernize my ERP first?
Both—ideally in parallel. A modern ERP creates the clean data foundation AI needs to deliver measurable value. Delaying either limits the return on both.
Can AI work on top of a legacy ERP?
It can run on legacy systems, but results will be unreliable. AI built on fragmented, outdated data produces what McKinsey & Company's Joe Boden calls "faster garbage out." Decision quality erodes quickly—and trust in the system follows.
How much faster is AI-enabled ERP transformation?
McKinsey & Company estimates up to 50% faster, with up to 60% fewer resources, compared to traditional approaches. Timelines still depend on scope, data readiness, and change management.
What's the biggest mistake companies make when scaling AI?
Treating AI as an IT experiment rather than a business transformation. Pilots succeed on hand-curated, controlled data—and fail when confronted with the full complexity of real enterprise environments.
What does "headless ERP" mean?
An ERP that operates as a backbone of data and business logic without a traditional user interface. Users interact through AI agents, natural language, or proactive notifications—not navigating screens.
solid-light
blue-1
blue-7
horizontal
light
blue
SAP logo

SAP Product

AI-Powered ERP for mid-market

Build the ERP foundation your AI strategy depends on—fast implementation, AI built in.

manual

Explore SAP GROW

link-target-same

secondary