Most Companies Get AI Adoption Backwards
Seventy-eight percent of organizations now use AI in at least one business function, according to McKinsey's 2025 Global Survey on AI. For most of them, that means scattered tool subscriptions, disconnected experiments, and a tips-for-prompting PDF from IT.
A smaller group is rebuilding operations from the ground up, treating AI as the operating system instead of an add-on.
The Core Problem
Most knowledge workers spend the majority of their time on execution: formatting documents, researching, checking details, copying and pasting between systems. Strategic thinking gets whatever time remains.
AI excels at execution. It cannot set direction, but once direction is set, it executes faster and more thoroughly than any human. Building an AI-native operation means deliberately flipping that ratio - freeing people to do judgment-driven work while AI handles the mechanical tasks.
One Platform, Not Twenty Tools
The first practical decision is choosing a single platform where all AI agents, skills, and institutional knowledge live. This prevents the fragmentation that kills most AI initiatives: different teams using different tools, knowledge walking out the door when people leave, and no learning that compounds across the organization.
Notion paired with Claude offers one model. Notion is already where work gets decided and tracked. It's model-agnostic, meaning the underlying AI can be swapped as the market changes without rebuilding the system. And it's accessible - every workflow is written in plain English and visible to anyone.
The concept of a skill is central. A skill is a set of refined, tested, plain-English instructions that captures how the best practitioners approach a specific task. When an SEO analyst needs to run a technical audit, they open Claude, describe what they need, and Claude identifies the relevant skill from the company-wide library, follows every step, and produces the deliverable. The analyst reviews it, applies judgment, and ships it. Every improvement to that skill immediately becomes available to everyone.
The Architecture Underneath
A central agent watches a task board inside Notion. When a new task appears, it reads the task properties and routes it to the right executor.
Simple administrative work completes itself without human or AI involvement. Tasks requiring external integrations - creating a Slack channel, sending a calendar invite - are handled by workers that connect to outside platforms via APIs. Complex work requiring AI and organizational knowledge calls Claude with the appropriate skill. Work requiring human judgment surfaces in Claude Cowork as assigned sessions, where the person provides direction and Claude handles execution.
Every action is logged. Every agent execution leaves a trace. Version history is preserved.
What This Actually Requires
A significant risk: the person leading AI transformation doesn't understand the tools at a practitioner level. Delegating to a consultancy or dropping it on the technology team alone tends to fail. The people who define what good output looks like are the practitioners doing the work, not the people setting up the infrastructure.
Companies should run structured training sessions across every region. Rather than product demos, make them hands-on. People build skills for tasks they actually do every week. By the end, participants should configure their personal instructions and build at least one working skill from scratch.
Some companies are also reframing roles around the new reality. Positions previously titled analyst or strategist are being reclassified as agent orchestrators and agent architects, reflecting what the work now actually involves.
The Phases
There is no universal timeline. A 50-person company could move through these phases in weeks. A company of several thousand might take a year. The sequence matters more than the speed.
Phase one: Organize the data layer. Before anything can be built, a company's knowledge needs to live somewhere structured and accessible. Processes, client information, policies, playbooks - all in one system of record. If knowledge is scattered across drives, wikis, and people's heads, AI has nothing to work with. This is the unglamorous phase most companies skip.
Phase two: Select tooling and build the first skills. The goal is not perfection but proving the pattern works in your specific environment. Build skills for the highest-volume, most repetitive work first. Get early adopters using them daily, measure time saved, and refine.
Phase three: Roll out training and scale. Once the pattern is proven, the wider organization needs hands-on training where people build skills for their own work. Governance gets established so skills are reviewed, quality-controlled, and shared across teams.
Phase four: Integrate and automate. The AI layer connects to core systems: CRM, project management, client platforms, internal tools. Work flows through the system with less human intervention at the routing level. Feedback loops tighten. At this point, a company is operating as AI-native.
The Timeline for Results
The compound effect of these efforts may take quarters to show up in performance data. That is the honest timeline. Any company serious about becoming AI-native needs to build with that in mind.
For operations professionals implementing these changes, understanding AI Agents & Automation and the practical skills required is critical. An AI Learning Path for Operations Managers can provide structured guidance through the phases of implementation.
Your membership also unlocks:
