Breaking through the "Non-Adoption" Barrier in Generative AI Implementation: Transforming Organizations in 5 Steps Using the ADKAR Model

Most failures in Generative AI implementation stem not from technical shortcomings, but from "psychological resistance on the front lines." No matter how sophisticated the LLM (Large Language Model) you build, if employees harbor anxieties that "their jobs will be taken away" or "their performance evaluations will suffer because they don't know how to use it," the system will become a mere formality, and the return on investment (ROI) will be zero.

This is where the "ADKAR Model", which frames individual mindset shifts in five stages, proves effective. In this article, we will explain strategies from a practitioner's perspective on using change management methodologies to dissolve resistance on the ground and ensure systems are adopted and used voluntarily.

1. [Awareness] Lack of "Awareness (Why do it now?)" leading to frontline rejection of AI

The most common situation we encounter in actual consulting engagements is when tools are introduced solely by executive decree, leaving the frontline staff without an understanding of "why AI is necessary now." If the first stage of the ADKAR model, "Awareness," is missing, the frontline will perceive AI as a "monitoring tool" or a "prelude to downsizing."

In our consulting engagements, we first use data to demonstrate that existing workloads are nearing their limits, and begin by presenting AI as a tool for 'augmentation' rather than 'replacement,' alongside specific use cases. It is particularly important to translate the message 'Why generative AI is essential to maintaining our company's competitiveness' into the language of the frontline.

2. [Desire] Designing "Desire" (the motivation to act) to foster psychological safety

The resistance of "I understand the necessity, but I don't want to do it" is the barrier of the second stage, "Desire." At this stage, it is necessary to clarify the "WIIFM (What’s In It For Me)"—how using AI will change individual evaluations and how much it will liberate them from tedious routine tasks.

To manage psychological resistance, providing a "sandbox" where failure is permitted is also effective. Foster a culture that encourages small "trials," such as starting by having AI generate draft responses for customer support occurring in the field of In-house EC Site Construction and Growth Support.

3. [Knowledge & Ability] Creating a Success Loop with "Knowledge" and "Ability"

Not knowing how to use something (Knowledge) and being unable to master it (Ability) are fundamentally different. Simply distributing manuals is insufficient; hands-on training tailored to the specific operational context is essential.

To enhance Ability (Execution Capability), it is effective to standardize prompts through templates and establish a proprietary internal RAG (Retrieval-Augmented Generation) environment. The key to preventing these initiatives from becoming a mere formality lies in a design that reduces the burden on frontline staff to "think from scratch," shifting their role to "verifying and refining" the AI-generated output.

4. [Reinforcement] Breaking through barriers to adoption: "Reinforcement"

The final phase, "Reinforcement," is the one most often overlooked. Three months after implementation, many users tend to revert to their old, familiar ways. To prevent this, it is essential to have a system in place to visualize the time saved and value created through AI utilization and to publicly celebrate these successes.

In our consulting practice, we recommend institutional approaches such as sharing AI success stories company-wide on a weekly basis and incorporating "Operational Improvement via AI" into performance evaluation criteria. Reforming organizational culture is not a one-time event; it is the process of rewriting everyday evaluation standards.