How a 17th century urban planning lesson can save your AI strategy
Imagine the Boston of 1630. A young Puritan colony sprawling on a rocky peninsula, where roads do not yet exist and cattle roam freely through meadows and hills. The cows, with that pragmatic animal wisdom of theirs, trace natural paths following the path of least resistance: bypassing boulders, avoiding swamps, connecting pastures and watering holes.
Decades later, when the city's founding fathers were faced with the need to create a road system, they made a decision that seemed reasonable: instead of designing a logical and orderly grid from scratch, they simply paved the paths already laid out by livestock.
The result? The chaotic maze of winding streets that still characterizes downtown Boston, where Washington Street meanders like a river gone mad and where even the most sophisticated GPSs sometimes give up in frustration.
Historical source: The story is documented in the poem "The Calf-Path" by Sam Walter Foss (1858-1911), which tells precisely how the paths traced by a calf later became the streets of a city.¹
The Boston story is fascinating because it perfectly illustrates a paradox: what works locally and immediately can prove disastrous on a larger scale and in the long term. The cows were right to follow the path of least resistance for their immediate purposes, but their paths were not designed for carts, cars, trucks, or city buses.
The lesson is profound: not everything that develops organically is optimal for the future.
On modern farms, "cow paths" are everywhere. They are those processes that have developed organically over time. As Jim Highsmith explains:* "In the IT world, 'paving cow paths' means automating a business process as it is, without thinking too much about its effectiveness or efficiency"².
These processes were formed like cow paths: following the path of least resistance at the specific time they were born. But now, in the digital age, continuing to follow them can be devastating.
When companies decide to "digitize," they often fall into the same trap as Boston's founding fathers. They take existing processes and "pave" them with technology:
"Have we always filled out this form by hand? Perfect, let's create a fillable PDF!"
This is digitization: converting analog to digital without changing anything substantial. As Gartner defines, "Digitization is the process of converting analog information into digital format."³ It is like paving the cow path - it becomes smoother, but remains tortuous and inefficient.
A manufacturing company I know had a quality control process that required 14 different steps, developed gradually in the 1980s and 1990s. When they "digitized," they simply transferred all 14 steps to tablets. The process became faster, but it remained fundamentally irrational: 8 of those steps were duplicate or obsolete.
True digitization means doing what Boston should have done: looking at the end goal and designing from the ground up the best way to get there.
According to the Gartner glossary, "Digitization is the use of digital technologies to change a business model and provide new opportunities for value and revenue; it is the process of moving to a digital business."⁴
Examples of actual digitization:
Today we are seeing a new wave of "paving the cow path" with artificialintelligence. Companies are taking inefficient processes and enhancing them with AI, creating what we might call "super-lasticity."
As highlighted in the Harvard Business Review, "The idea of business process reengineering is coming back, this time driven by artificial intelligence. In the 1990s, the implementation of ERP systems and the Internet enabled changes in business processes, but expectations of radical change were often not met. However, AI enables better, faster, automated decisions."⁵.
Inefficient processes that are now inefficient faster and more accurately.
Before implementing any technology, follow this order according to Michael Hammer⁶ methodology:
Obliterate: Eliminate anything that does not add real value
Integrate: Connect remaining processes into logical flows.
Automate: Only at the end, apply the technology
As Hammer writes, "It is time to stop paving cow paths. Instead of incorporating obsolete processes into silicon and software, we should obliterate them and start anew."⁷
These terms, borrowed from urban planning and software engineering, define two radically different approaches⁸:
Brownfield (Paving the path):
Greenfield (Design from scratch):
As highlighted by McKinsey, "While 90 percent of companies have begun some form of digital transformation, only one-third of the expected revenue benefits have been realized"⁹.
Before any technological implementation, ask:
One European bank had a loan approval process that took 45 days and 12 different steps. Instead of "digitizing" the existing process, they completely redesigned it:
The secret? They realized that 90% of the controls were redundant and that AI could assess risk more accurately than 6 different offices.
One Italian hospital had waiting times of 4 hours in the emergency room. Instead of "digitizing" the queuing system, they completely rethought patient flow:
Result: Waiting time reduced by 80%, patient satisfaction +60%.
"Our employees are used to this" is the most insidious killer of innovation. It is like saying that cows were used to their own paths.
"We have already invested so much in this system" ignores the fact that continuing on a wrong path amplifies the error.
"It's too complicated to change everything" often hides a fear of admitting that the current process does not make sense.
Don't start with the technology, start with the business objective.
Ask yourself, "If I were a company born today, how would I solve this problem?"
Redesign processes for the digital age, not translate them into digital.
Using an incremental approach but with a radical vision.
Measure not only efficiency, but overall effectiveness.
The process of transformation never ends.
Be careful that new spontaneous "cow paths" do not form.
Artificial intelligence can be either the best tool for paving cow paths (making them super-efficient but fundamentally wrong) or the best architect for designing the cities of the future.
As the World Economic Forum subtends, "To reach its full potential, AI needs to speak the language of business, it needs to understand how work flows, and it needs process intelligence"¹⁰.
The difference lies in the approach:
The deeper lesson of the Boston story is not technical, but psychological: it takes courage to admit that the paths we follow are not necessarily the best possible ones.
In business, this means:
Today, faced with the endless possibilities of AI and digitization, we have a choice: we can do as Boston's founding fathers did and pave existing paths, or we can have the courage to design the cities of the future.
The next time you hear the phrase "let's digitize this process," stop and ask, "Are we designing a modern road or are we paving a cow path?"
The future belongs to those who have the courage to leave the beaten path and design new roads. Even if this means admitting that cows, wise as they were, were not urban planners.
"It is time to stop paving cow paths. Instead of incorporating obsolete processes into silicon and software, we should obliterate them and start over. We should 'reengineer' our companies: use the power of modern information technology to radically redesign our business processes to achieve dramatic improvements in their performance." - Michael Hammer, Harvard Business Review, 1990¹¹
.svg)
.svg)
.svg)
.jpeg)
.jpeg)