We've seen this movie before

AI is giving us a chance to redesign how work actually gets done. Most companies are using it to go faster in the wrong direction.

Reminds me of the Business Process Re-engineering push. Same failure mode, different decade. Hammer and Champy warned us that automating a broken process just gives you a faster broken process. The unlock isn't better process mining. It's designing new processes around what AI actually makes possible.

I've been watching a familiar pattern emerge across the contact center and customer service space. It's one I've seen before. The first time around companies were deploying recording software to capture every click and keystroke their agents make, then wrapping AI agents around those exact workflows. The pitch is transformation. The reality is acceleration of the status quo.

This is Business Process Re-engineering 2.0. And we're already making the same mistakes.

What BPR got right — and where it went wrong

In the early 1990s, Michael Hammer and James Champy published Reengineering the Corporation and sparked a wave of operational transformation across industries. The core insight was radical for its time: don't automate existing processes, redesign them from scratch around the capabilities of new technology.

The companies that did it well — genuinely rethought their workflows rather than digitized their org charts — captured enormous gains. The ones that failed took their existing processes, drew some boxes around them in Visio, and called it reengineering. They just made the old approach run faster.

"Automating a broken process just gives you a faster broken process. The unlock isn't better process mining. It's designing new processes around what AI actually makes possible."

Sound familiar? It should. Because that's the dominant mode of AI deployment I see today.

The contact center is a perfect case study

Consider what most enterprise contact center architecture actually looks like. Decades of workflow were built around the constraints of the telephone: IVR menus, hold queues, escalation trees, after-call work, quality scoring rubrics. Every one of those structures was a workaround for the limitations of the technology available at the time.

Now AI vendors are walking into those environments and offering to automate them. And enterprises are buying it — because it's measurable, low-risk, and maps neatly onto existing ROI frameworks. But here's the question that rarely gets asked: why does this workflow exist at all?

The IVR was built because routing a call to the right person was expensive and slow. If reasoning can be built into the system from the start and if an AI can understand intent, access context, and orchestrate outcomes in real time then the question isn't how to put a voice agent in front of the IVR. The question is whether the IVR should exist.

The companies getting real ROI are asking different questions

The AI deployments generating genuine transformation share a common characteristic: they started with a blank page, not a process map. Rather than asking "how do we automate this?" they asked "what should this actually look like now that reasoning, context, and orchestration can be built into the system from the start?"

That forces a different conversation. Not just how to speed up the current workflow, but what should be simplified, eliminated, kept human, or redesigned entirely. It's a harder conversation. It requires rethinking incentive structures, org design, and vendor relationships — not just deploying a new tool.

But it's the only conversation that leads to real transformation.

What BPR 2.0 actually requires

The original BPR movement gave us a useful framework, even if its execution was sometimes botched. The principle holds: the right question is not how to automate the current state, but what the right state looks like given new capabilities.

For AI vendors selling into the enterprise, especially in contact centers and customer service, this has real implications. The companies that will win in this market are not the ones with the best process mining tools. They're the ones helping enterprises ask the harder question: what does a customer interaction actually look like when you design it from scratch around AI?

Not mining the current state more precisely. Not recording more clicks. Not agentifying yesterday's operating model.

The real value is in redesigning the work itself. We've had the tools to ask that question before. This time, we actually have the technology to answer it.

AI strategy Contact center BPR 2.0 Enterprise AI