The Illusion Of Collaboration.
Exploring why messy friction in teamwork is our edge in an AI-smoothed workplace.
You are in a meeting. Everyone nods in sync.
The AI notetaker – because there is one, admit it, since the last time you could not decipher your own handwriting – captures every word. The auto-summary highlights next steps. Thanks to efficient workflow automation, a draft email is already waiting in your inbox.
Flawless. Frictionless. Efficient. Very much in line with what many companies obsess over today.
Here is the thing: collaboration is not supposed to be flawless, frictionless, or efficient. It is supposed to be messy.
Here is another thing: disagreements, misunderstandings, and awkward silences are not inefficiencies. They lead to progress.
With its polished responses, an AI chatbot is making teamwork smoother. Isn’t it making it shallower too?
Friction is fuel
Let me be clear upfront. Personal spats are not included in my definition of friction. Relationship issues in the workplace have always existed. Most of them are emotionally draining for the parties involved, and value-destructive for the business.
Similarly, positive forms of conflict or tension have always been present at work. My hypothesis is that they are fundamental for us to reach the ideal human-AI balance in the workplace, marrying efficiency gains from AI to greater meaning for humans.
Recent experiments tend to confirm my hypothesis — although I will humbly say there is still a lot of uncertainty as to what an AI-augmented workplace will look like in a few years.
One experiment showed that while AI tools can boost speed and reduce friction in collaboration, they can also limit diversity and richness of outcomes when creative tasks are involved. For example, in human-AI marketing teams, communication surged and repetitive editing declined, but human teams still outperformed in image creativity. Another experiment showed that when people use many AI-generated prompts, there’s more idea diversity across the group, but individual creativity risks conforming to patterns.
That challenge, that struggle, that push-back is what often leads to breakthroughs. The iPhone may still have a physical keyboard if there had not been friction between Steve Jobs and his engineering team. Constructive friction was even institutionalised at Pixar.
As long as all parties have a collective desire to move forward together, frictions are a positive thing.
The workplace is not a binary, win-or-lose theatre. It is made of multiple shades.
Collaboration without effort is coordination, not co-creation
Resorting to chatbots to smooth out differences in these shades can achieve harmony and speed, but it depletes the creative collaboration process. By doing so, you end up with a me-too product, a bland presentation, or a neutral customer experience, depending on what you are building.
As mind-blowing as frontier models are, they also represent a path to mediocrity. There is no shortcut to expertise. There is no shortcut to greatness. And guess what, there is no shortcut to efficiency. The more frictionless it looks, the deeper the thinking has been, the more effort there has been. Time, effort, and persistent challenge — what some call ‘grit’ — consistently produce better results than shortcuts, whether in Pixar films, Apple devices, or deliberate practice studies.
I once read in Sahil Bloom’s excellent newsletter a phrase that resonated with me: “embrace the suck.” In this context, it is not always pleasant to have a work-related argument with colleagues. In certain cultures, it is even to be avoided at all costs. But when done right, temporary antagonism always elevates the outcome. That’s not all: it teaches the protagonists how to rise to a challenge and even solidifies the relationship between them. And no technology was involved to elevate us here.
As AI continues to make strides, the world of work is undoubtedly going to evolve yet again.
Except that this wave of change is bringing a fundamental question: what is work?
The human edge is in the mess
If you haven’t read anywhere that “AI is coming for your job” or “AI will replace you,” I assume you’ve been living under a rock. Chances are that you are not if you are reading this newsletter.
Rest assured. AI is still coming for your job. But you’ll have a better one.
As humans, we are messy. We achieve levels of thought randomness that LLMs can’t approach yet. Humans bring a randomness, a serendipitous connection-making, that AI still cannot replicate — a source of creativity that drives breakthroughs in ways machines struggle to emulate.
And in that randomness, we connect paths that were either unseen or simply inconceivable. This creative process of assembling seemingly disconnected pieces is hard to replicate, as it sometimes seems to come out of nowhere. This process is magnified when done in like-minded teams. In those instances, we perform combinatorial creativity, and today’s AI just can’t compete.
It may look far-fetched, naïve, or optimistic (or a combination of the three!), but I do see work in the future as a space where:
Workers devote their time to think, debate, and problem-solve collaboratively, to get to that spark that enlightens the path to a solution for a client’s problem.
Technology — namely AI in different flavours, from robotics to GenAI — is pervasively present and yet invisible, doing what it does best: responding to increasingly complex prompts to generate increasingly refined content.
An AI-infused workplace is not a place where technology is over-present. On the contrary, if it is the case, tech becomes an unnecessary centre of attention, distracting people from real collaboration.
The place of AI within businesses and their intrinsic collaboration mechanics is still unknown. Unquestionably, AI will take over many of the routine tasks humans were executing. Undoubtedly, AI will be elevated to the level of a virtual colleague in specific instances.
The way jobs, roles, and businesses are currently designed won’t make much sense anymore as AI permeates our lives. At work, AI will clearly lift thinking constraints, operating either as a sounding board or an execution planner. This looks promising, but it forces us to adjust our own ways of working.
As always with potentially disruptive changes, it’s about anticipating and accepting the implications.