The Bus from Milan: The Framework Survives the AI Shift
Thinking in Motion: Article 7 of 7 Category: Thinking in Motion (/thinking-in-motion/) Site: jasonbarnard.com
Fourteen Hours, Two Seats, No Internet, and the First Constraint I Chose Deliberately
Milan to Montpellier, fourteen hours by bus, and I bought two seats, because by 2024 I had learned a few lessons about how helpful long haul travel can be.
The train in 2012 had taken my money and given me a methodology, the car in 2017 had taken the radio and given me a framework, the flights had taken my sleep and my laptop and given me a universal theory, and every breakthrough had followed the same pattern (something removed, constraint imposed, synthesis in the space that opened up), and this time, for the first time, I chose the constraint deliberately rather than having it imposed on me.
Two seats meant no stranger sitting next to me, no small talk, no interruption, and a bus rather than a plane meant more time in the condition that produces the thinking, and no internet meant fourteen hours of nothing except what was already in my head.
The Landscape Was Shifting Beneath Everything I Had Built
I had been solving algorithms for twelve years by this point, built a platform, processed billions of data points, and the methodology was proven, the evidence base was deep, clients were seeing results, and the industry was slowly catching up to ideas I had planted a decade earlier, but I could feel that the landscape was shifting beneath everything in a way that made the previous shifts look minor.
AI was no longer a feature inside search engines but was becoming the interface itself. ChatGPT, Perplexity, Claude, Copilot, Gemini, each one a new system that needed to understand entities, evaluate credibility, and make recommendations, and each one following the same fundamental pipeline I had mapped, but with different mechanics, different training data, different presentation formats, and the question I took onto that bus was the one that mattered most: does The Kalicube Process™ survive the transition from search engines to AI, or is it a search-era methodology that breaks when the interface changes?
Why AI Makes Algorithmic Identity More Important, Not Less
Fourteen hours later I had my answer, and the answer was that the methodology does not just survive the transition but becomes dramatically more important because of it.
In traditional search, the algorithm presents ten links and the user decides, which means the Brand SERP is a shop window, and if the window is wrong you lose the click but the user still has options because they can scroll, click elsewhere, refine their search, and eventually find you through a different path. In an AI-driven interface, the algorithm does not present options but makes a recommendation, stating with its full authority “based on what I know and trust, this is the answer,” and there is one answer, not ten links, and there is no second page and no “scroll down and find the alternative,” because the AI decides and the user follows.
This means the stakes of algorithmic identity are now higher than they have ever been, because a wrong Brand SERP cost me a hundred thousand euros in 2012, but a wrong AI recommendation costs you every potential customer who asks that AI for advice in your category, twenty-four hours a day, seven days a week, across every platform that runs the model, and that is not a single lost deal but a permanent leak in your revenue that compounds with every query.
Seven AI Platforms Are Your Untrained Salesforce, Working Every Hour of Every Day
The UCD Framework maps directly onto this new reality (Understandability: does the AI know who you are and what you do? Credibility: does the AI trust the information enough to recommend you? Deliverability: will the AI present you to the right user at the right moment?), and the stages have not changed since I first sketched them in those Metz slides in 2015, but the consequence of failure at each stage has multiplied by orders of magnitude.
I had been calling this “Answer Engine Optimisation” since 2017, when I coined the term, but on that bus I understood that even that framing was too narrow, because these are not answer engines and they are not search engines but are AI assistive agents, digital employees, each one making decisions about your brand on your behalf (or on behalf of your competitors) every second of every day.
This is what I later formalised as the Untrained Salesforce: seven AI platforms (Google, ChatGPT, Perplexity, Claude, Copilot, Gemini, and Siri) are your digital sales team, working twenty-four hours a day, seven days a week, three hundred and sixty-five days a year, never taking a holiday and never calling in sick, and they either sell for you or they sell for your competitors depending entirely on how well you have trained them to understand and trust your brand.
The Kalicube Process is how you train them.
The Pattern That Runs Through Every Transit Breakthrough Since 2012
The bus gave me what I needed, which was not a new framework but the confidence that the existing framework was built on principles deep enough to survive the biggest platform shift in the history of digital marketing, because the pipeline (discover, understand, evaluate, present) does not change when the interface changes, and the signals that feed the pipeline (consistency, corroboration, clarity) do not change because the AI is generative rather than retrieving, and the methodology holds because it was never about search engines but was always about how algorithms process identity, and algorithms will always need to process identity regardless of the interface they wear.
In 2012, a train took away a hundred thousand euros and forced me to confront the problem. In 2015, the snow in Metz revealed that the solution was already formed and waiting for words. In 2017, a broken car radio and a friend who sang for eight hours gave the framework its architecture. In 2019 and 2020, an investigation across four continents gave it evidence, a flight from Sydney gave it convergent proof, and a confiscated laptop over the Atlantic gave it universality. And in 2024, a bus from Milan confirmed that it holds, not just for search but for AI and for whatever comes next, because the principle underneath everything (tell the truth, tell it clearly, tell it consistently, make sure every source corroborates every other source) is not a technique but is how information systems have always worked and how they always will.
Every breakthrough required something to be removed: the money, the radio, the sleep, the laptop, the internet. Each removal was a constraint that forced a synthesis, and by the seventh transit breakthrough I had learned to choose the constraint deliberately, because the pattern was too consistent to be coincidence and too productive to leave to chance.
I bought two seats on a fourteen-hour bus, turned off the internet, and let the thinking do what thinking does when there is nothing else competing for the space.
The framework holds.
Thinking in Motion Series
Planes, Trains and Automobiles: Why Every Breakthrough Happened Between Somewhere and Somewhere Else
- The Train from Paris. Where Brand SERP was born (2012)
- The Slow Snow Train to Metz. Everything was already in the slides (2015)
- The Car to Metz. Hugo sang and the framework connected (2017)
- Twenty-Two Flights Around the World. The investigation nobody else did (2019)
- The Plane from Sydney. When the engineers confirmed everything (2019)
- The Flight from Seattle. Universal theory without a screen (2020)
- The Bus from Milan. The framework survives the AI shift (2024)
