When Humans Become the Long Pole in the ERP Tent

For years, ERP projects had some usual suspects when timelines slipped. Configuration took longer than expected. Integrations turned out uglier than the slide deck promised. Data migration became its own side quest. Testing found things no one mentioned in discovery. Everybody nodded like this was just the natural order of the ERP universe.

Now AI is starting to mess with that old pattern.

Suddenly, we can draft requirements faster. Generate test cases faster. Accelerate code reviews faster. Summarize workshops faster. Map processes faster. Clean and classify data faster. Build copilots, agents, and automations faster. We are entering a world where parts of implementation and transformation that used to crawl can now move with real speed.

That sounds like great news. It is. But it also raises a more uncomfortable question that does not get enough airtime in the ERP space:

What happens when the software is no longer the slowest part of the transformation?

What happens when the business becomes the bottleneck?

That is the question.

And no, this is not one of those “people are resistant to change” throwaway talking points people use when they do not want to think harder. This is bigger than that. Current research points to a messy middle: McKinsey says employee readiness for AI is relatively high and that leaders are often the real barrier to scaling it, while BCG reports that only 5% of companies are achieving AI value at scale and 60% are not seeing material value at all. In other words, faster technology does not automatically produce faster organizational absorption.

We used to wait on the system. Soon the system may be waiting on us.

In a traditional ERP implementation, the long pole in the tent was often technical. That does not mean business engagement was easy, but more often than not, the calendar got dominated by data conversion cycles, dev backlogs, integration specs, environment issues, and testing defects.

AI is beginning to compress parts of that work.

Not all of it. Not perfectly. Not safely in every scenario. But enough of it that leaders should start preparing for a different type of delay. Less “the build is not ready,” and more “the business has not decided,” “the team has not adapted,” or “the process owner still does not trust the new way.”

That is where this gets interesting.

Because most businesses are not actually structured to absorb change at AI speed.

They are structured to debate, escalate, revisit, and socialize change at human committee speed.

That worked, or at least sort of worked, when the technology itself moved slower. But when the technical side starts accelerating, every crack in your operating model becomes painfully obvious. McKinsey has made a similar point in digital and AI transformations: the issue is often not the technology itself, but organizing talent, data, and the business around it, and technology tends to expose the next operational bottleneck in the chain.

The dirty little secret: speed creates more change, not just less work

This is the part many people miss.

When AI speeds up delivery, it does not just reduce effort. It increases the rate at which the business must absorb change.

That means more decisions per week. More process redesign per phase.More frontline adjustment in a shorter window. More governance touchpoints. More frequent training.More need for champions. More management visibility. More pressure on leaders to make tradeoffs clearly and quickly.

More does not always equal more.

So what should businesses do about it?

First, stop measuring readiness like it is a soft side issue.

If AI is accelerating implementation, then organizational change capacity needs to be treated like a hard dependency. Not a side workstream. Not a nice-to-have. A dependency.

Second, start putting decision SLAs on the business, not just tasks on the SI or IT team.

If the future bottleneck is human decision throughput, then leaders need to govern it like they govern budget and scope. You do not get to demand a faster program while taking two weeks to confirm a process owner.

Third, shrink the blast radius.

Do not try to transform every role at once just because the tools make it possible. Sequence the change. Pick the highest-friction, highest-value workflows first. Let the organization build confidence before you ask it to absorb a whole new operating model.

Fourth, measure adoption in behavior, not attendance.

Do not tell me 200 people attended training. Tell me whether planners changed how they run supply review. Tell me whether buyers stopped using side spreadsheets. Tell me whether warehouse leads follow the recommended exception path. Tell me whether finance trusts the close analysis enough to shorten review cycles.

That is adoption.

Fifth, design every AI-enabled ERP initiative with four owners in mind

1. business process owner

2. technical owner

3. change owner

4. risk owner

If one of those is fuzzy, the value will be fuzzy too.

Microsoft’s own adoption playbooks are pretty blunt on this. They call for an executive sponsor, champions and early adopters, user enablement, a learning community, technical readiness, and ongoing usage monitoring. In other words, even the vendors pushing AI hardest are not pretending this is just a deployment exercise.

The real competitive advantage may be change absorption

This is where I think the market is heading.

In the last era, competitive advantage often came from who could implement the best system.

In the next era, it may come from who can absorb the fastest rate of meaningful change without organizational whiplash.

That is a different muscle.

• It is leadership clarity.

• Process ownership.

• Governance discipline. Training maturity. Trust-building.

• Role redesign.

• Decision speed.

• Manager capability.

AI may absolutely make ERP implementations faster. I believe it will.

But the companies that win will not just be the ones with the best prompts, best copilots, or best agents.

They will be the ones whose people, leaders, and operating model can keep up with what the technology now makes possible.

That is the question more ERP leaders need to start asking in every steering committee, every transformation kickoff, and every AI strategy session:

Not “Can we automate this?”

But “Can our business actually absorb what automation is about to accelerate?”

Because if the answer is no, the next bottleneck is not your system.

It is you.

And that is not an insult. It is just the new project plan.

Dad joke before I go: buying a faster lawn mower does not matter much if the kid pushing it is still negotiating the meaning of “done.”

Stay curious, stay honest, and do not mistake technical acceleration for organizational readiness.

DynamicsDad