At Monki Gras 2026, Laura Tacho called out that what holds us back are our human and systems-level constraints. Not the technology - Us and the ways in which we organize and communicate (or don’t). 

We’ve been building software to avoid talking to each other for decades—user stories without a user and missing APIs, incomplete documentation that lives in someone’s head, onboarding that is “just ask someone.” These aren’t technical failures. They’re communication failures and AI excavates and puts a spotlight on them. When you ask an agent to operate in your environment, it needs what your team never fully codified: the context of why a decision was made, the shared language of your domain, the goal behind the request—not just the request itself. The “stuff” we never did well for our human colleagues is missing in what we feed the machine.

We’ve seen this movie before.

Agile started as a movement about people and interactions over processes and tools. At its core DevOps is about breaking down the wall between development and operations - a change in methodology and culture that enabled teams to actually release software faster and more reliably. The teams that did this successfully always shared stories about the transformation in how their teams worked together vs. what products they used. Technology was the enabler to the new model of work. Over time they became product categories and gave organizations an easy way to buy a product to solve organizational problems (which will just exacerbate your problems). 

Now AI makes a similar promise: agents will collaborate so humans don’t have to. Which sounds compelling until you realize that the friction you’re automating around is often the thing that built alignment and trust in the first place. Matt LeMay’s Monki Gras talk made the point that planning is communication, and communication is craft. When we treat it as overhead — something to be minimized, replaced with a status update in a tool — we erode the thing that makes systems resilient. Trust is not a feature you can ship.

Building value in a business requires a group of people to be aligned to the same goal – delivering products or services that customers want and the market recognizes and useful. A shared understanding of the goal (outcome), language/definitions, context (data, history), and agreed upon methods for ongoing communication and synchronization. Funny enough, the machine needs the same thing. 

Where teams break down falls into three buckets: communication that is absent or noisy; understanding -  misaligned priorities with no shared definition of success; and feelings - the interpretations we make when information is absent, which attack under pressure. These problems also break AI adoption. Technical debt gets the attention, but cognitive debt is a silent killer — the accumulated weight of decisions and processes never documented and context that exists only in the memory of people who may leave tomorrow — is the one that compounds invisibly. When the why behind a decision is lost to a Slack thread from three years ago, AI literally cannot proceed without the information you never bothered to record. It’s not a limitation of the model but a mirror. No amount of prompting can fill the gap of missing knowledge and context. 

The space between is where the real work is.

The gap between what AI can do and what organizations are actually able to absorb isn’t a technical gap. It’s a human one - the communication that didn’t happen, the missing information, the assumptions, and the trust that was skipped in service of speed. The answer isn’t to simply replace humans with AI or avoid AI. The space between people is the same as the one between people and machines. Between intent, instruction, and execution. To do that well, we need well intentioned people to participate in its definition, curation, and use. 

Early computers spoke their own language to help humans run more calculations, faster, and more accurately. Grace Hopper believed the real limitation was that computers couldn’t communicate – leading to the development of compilers and programming languages. A huge step forward in the machine to human communication but limited to those who have the fluency in these specialized languages. Today’s innovation goes further to make communication between humans and machines more accessible. With that, we need to think about how we redesign what we do and how we do it — with tech as an extension of us — with clear intent and instruction — to deliver what we expected.

Danilo Campos’ talk from Monki Gras sums this up for me. We are at an inflection point for cognitive tooling—the pace is no longer linear, and the infrastructure hasn’t caught up to what’s possible. For example: it took 68 years to reach the first terawatt of solar power capacity, two years for the second. Moments of abundance tend to reward those who were already doing the underlying work. If you’ve been building systems with clear documentation, explicit context, and strong communication norms, AI amplifies that. If you’ve been deferring it, hoping tooling would eventually solve it, AI exposes that—faster and more visibly than anything before.