AI: The Answer to a Business Process Re-engineer's Prayer

Tim Baker

A converted sceptic with 40 years of scar tissue

TCB Consulting Ltd | October 2026

For thirty years, two stories have been running in parallel in the world of business technology. They started in different places, spoke different languages, and for a long time, seemed to have nothing to do with each other. One was a story of diagnosis. The other was a story of tools. Now, they have finally converged.

It began in the early 1990s with a radical idea: Business Process Re-engineering. The promise was seductive. Don’t just automate what you already do; obliterate it and start again. [1] Find the root cause of the inefficiency, the bottleneck, the delay, and redesign the entire process from a clean slate to serve the customer. It was a story about asking ‘why?’

At the same time, another story was unfolding in research labs and universities. It was a story about artificial intelligence, a field that had been through its own cycles of hype and disappointment. It was a story about building systems that could learn, reason, and understand. It was a story about creating tools, not just following instructions. It was a story about asking ‘how?’

BPR was brilliant at diagnosis. Consultants armed with whiteboards and process maps could brilliantly expose the hidden wiring of a company. They could show exactly where a process was broken, where the handoffs were failing, where the rework loops were burning time and money. They could design elegant, streamlined new workflows that were obviously, undeniably better. The diagrams were beautiful. [2]

AI, for its part, was brilliant at narrow tasks. It could play chess, recognise images, and translate languages. But it struggled to connect those tasks into a coherent whole. It was a collection of powerful but disconnected capabilities, a set of solutions in search of a sufficiently complex problem. It lacked a unifying purpose.

The problem for BPR was never the diagnosis; it was the cure. The beautiful diagrams were often impossible to implement. The cost of building the custom software, retraining the staff, and restructuring the departments to make the new process a reality was astronomical. The failure rate of BPR projects was notoriously high — often cited as 70% or more. [3] It was a discipline of perfect vision and limited power. It knew what needed to be done, but it couldn’t do it.

The breakthrough for AI came with the rise of the agentic model. Instead of a collection of narrow tools, we now have autonomous agents that can understand a goal, make a plan, and execute a multi-step process. They can interact with multiple systems, handle exceptions, and learn from experience. They are not just tools; they are digital workers. [4] They had the power, but they were waiting for a clear instruction.

And here is where the two stories meet. The prayer of the business process re-engineer has been answered. The implementation gap that plagued BPR for thirty years has been closed by the arrival of agentic AI. The ‘why’ has finally met the ‘how’.

Consider a classic BPR problem: a complex, manual, multi-stage approval process for a new client in a regulated firm. The BPR analysis identifies the bottlenecks. The redesigned process is a streamlined, single-touch workflow. In the 1990s, that was a multi-million-pound IT project. Today, it is a set of instructions for an AI agent:

Receive the initial application.
Perform the required background checks using external APIs.
Check the application against internal compliance rules.
Route it to the correct human for final approval if required.
Update all relevant internal systems.

This is the synthesis. The diagnostic power of BPR to identify the problem, and the implementation power of AI to solve it. It is not about replacing humans, but about finally freeing them from the tyranny of broken processes — the very thing BPR set out to do three decades ago.

The tools are here. The methodology has been proven. The only thing missing is the will to connect the two.

References

[1] Hammer, M. (1990). Reengineering Work: Don't Automate, Obliterate. Harvard Business Review.

[2] Davenport, T. H. (1993). Process Innovation: Reengineering Work through Information Technology. Harvard Business Press.

[3] Bashein, B. J., Markus, M. L., & Riley, P. (1994). Preconditions for BPR success. Journal of Management Information Systems, 11(1), 7-8.

[4] Wang, L., et al. (2024). A survey on Large Language Model based Autonomous Agents. arXiv preprint arXiv:2403.12293.