Most organizations think of productivity as a flat line on a spreadsheet. Maya knew it was a quiver. She remembers the townhall not for the speeches, but for the silence that followed—the gap between the leaders “inviting ideas” and the reality of how work actually gets done. While the company was drafting emails, Maya was building. Her prompts weren’t just code; they were a digital second skin. They sensed her tone, anticipated her fatigue, and did the heavy lifting of research while she did the thinking.
Maya was proud. She liked being faster than everyone else.
Then came the “Standardized Tooling.” In the name of uniform productivity, the organization tried to pave over the “Secret Garden” Maya had grown. The new tools were efficient, but they were brittle. They were built for the Median User—that mythical, average person who doesn’t exist.
Maya didn’t want a corporate interface; she wanted her mirror back. Her self-created tools understood her specific weaknesses and her private ambitions. They knew who she avoided and who she prioritized. When software is standardized, it doesn’t just lose its edge; it loses its soul.
The organizational chart is a lie. The real map of a company is drawn in the shadows.
Maya knew that a ticket isn’t just data; it’s a request from a human. She knew that Steve needed a phone call before an approval, and that certain teams were drowning and needed their queues skipped. This is the Shadow Process. It is the invisible nervous system of the office. If you remove Maya’s shadow, the machinery of the business doesn’t just slow down—it breaks. You cannot automate a process if you don’t understand the “shadow machinations” that actually make it work.
It’s Maya’s shadow that bridges the gap between processes and successful outcomes with its shadow machinations
The Organisational map has Maya’s shadow imprinted all over it.
We have spent decades building software that breaks the moment a human acts like a human. I call the antidote Plasticity. An agent that cannot mold itself is just a faster way to fail.
To build for the Mayas of the world, we need three things:
The Individual Mold: It must mirror the user’s specific “edge,” not the corporate average.
Vibration Absorption: It must sense when the “scaffolding” is shaking—knowing that if one team is late, the whole lattice is under tension.
Elastic Guardrails: Rules are necessary, but they shouldn’t be a chokehold. The guardrail defines the boundary; plasticity defines the dance within it.
For a builder, these aren’t just features; they are a shift in how we architect workflows. To build for Maya, we have to stop building for the ‘Median User’
Turn shadow to specialization
Design the agents to balance Enterprise processes, with Maya’s need for autonomy. The Agent should not be “averaging-out” Maya’s edge, but allow Maya to “upload” her excellence into her specific instance of the tool. Her “secret garden” isn’t paved over; it is digitally reinforced.
As Product builders of Enterprise workflows, we are often closer to Organisational processes than the users themselves.
Creating plastic agents mandates the need to go deeper, understand the users deeper, their psychology, problems, its nuances. Once you are able to create a map of how the expertise is shared between an Agent and a user, you will be able to define those interaction points better.
For example, prioritizing interviews for thousands of applications is never an easy job for a system. It includes a structured approach (a set of conditions like highest match, or first-in-first-out), but it also requires multiple context injections (who the Hiring Manager likes best, or who can go the fastest, or who has a competing offer).
Scheduling Interviews is often an art and experience. Only a human, interacting with same set of people know that the Jill is always busy and his time needs to be prioritized, or that Mason never shows up for any meeting before 10 am (even though his calendar is free).
Creating Plastic agents require extreme clarity on interaction points where context injection is valuable.
Absorb the vibrations of the scaffolding
Maya knows that when the sister team’s data is late, the whole scaffold starts to shake. She doesn’t wait for a system error; she senses the vibration and adjusts her pace. A plastic agent shouldn’t just record a failure; it should feel the tension in the lattice and adapt
Maya’s automations failed when the input data failed to follow a strict format. Or, the data was inconsistent or way out of range. It was Maya’s knowledge that filled this gap.
Agents, instead of Maya needs to absorb these vibrations in the Enterprise scaffolding. Agents need to be aware of these vibrations and know how to adapt per them.
Design guardrails cohesively with context injection
AI ethics and governance is being discussed in every Boardroom today.
Creating plastic Agents is not sufficient. Allowing that plasticity to meld well with the guardrails designed is a must.
The love for rigidity and adherence to processes often creates this myopia for Organisations where they do not acknowledge the need for plasticity. And that is where guardrails are designed to keep the shadows out of the Agents.
The guardrail defines the boundary, but plasticity defines the movement within it. A plastic agent knows that the rule is the guardrail, but the ‘Shadow’ is the reason the rule exists. Harmonize them by allowing the agent to flag when a rule is preventing a successful outcome.
As builders, harmonization of guardrails with its plasticity is a must.
If we continue to build agents that prioritize the “Official Process” over the “Shadow Map,” we aren’t building intelligence. We are building an expensive, automated bureaucracy.
By forcing Maya into a standardized box, we are effectively lobotomizing our most talented people. The goal of a Plastic Agent isn’t to make Maya follow the rules—it’s to give her a tool as brilliant, as flawed, and as effective as she is. We don’t need machines that act like machines. We need machines that finally learn how to breathe.