AI Renovation Initiative for the Entire Company: A Cowork Skill for Supporting the Creation of As-Is and To-Be Analyses

1. Introduction

At Classmethod Operations Co., Ltd., we are working on a project to incorporate AI and automation to transform our operations around AI (hereinafter referred to as "AI Transformation"). Within this initiative, our BGU (BPO Promotion Unit) provides accompanying support for organizing and improving operational processes across each department.

As shown in the figure, BGU plays a supporting role in AI Transformation across departments. The department promoters appointed by each department are responsible for creating the AsIs, which organizes current operations, and the ToBe, which designs improved operations.

BGU's first initiative was therefore to make it easier for department promoters to start organizing AsIs and creating ToBe.

This article introduces the current state of company-wide AI Transformation efforts and the two Cowork skills we have developed. It also describes the remaining challenges in the skills under development and our future improvement goals.

In relation to this article, we are scheduled to present at a DevelopersIO mini-session on 2026/10/24 under the title "Company-wide AI Transformation and Accompanying Support Using Cowork Skills."

2. Company-wide AI Transformation and Current Initiatives

In the AI Transformation, each department organizes its key operations, selects targets for improvement, and designs improved operations. Following that, prototyping and effectiveness measurement will be conducted, with plans to promote adoption into operations and expand to other business areas.

Methods for reviewing operations are not limited to replacement by AI. We refer to the ECRS concept of eliminating unnecessary tasks, combining similar tasks, changing the order or responsible parties, and simplifying tasks. Automation through AI or existing systems is a means chosen based on that review.

For such a review, it is first necessary to organize the current operations in a way that stakeholders can examine. After sharing who is doing what, what information is being used to make decisions, and where waiting or rework occurs, we think about the improved operations.

In terms of company-wide AI Transformation progress, we are at the stage of advancing the organization of current operations (AsIs) while beginning to enter the creation of the appropriate flow of our own operations (ToBe).

3. Background of Cowork Skill Development

When creating the skills, our first consideration was how to reduce the burden on department promoters.

We developed the Cowork skills because we believed that drawing a map on a blank sheet is the most difficult task.
The staff in each department understand the flow and procedures of their own work. However, even if they understand the operations, it is not easy to write out business flows and improvement proposals from scratch one by one. Moreover, the time available to work on this is limited when they are caught up in their daily duties.

If someone is not accustomed to organizing operations or designing improvements, they may be unsure of where to start writing and how much detail to include. We thought this could make it harder to take the first step.

Therefore, BGU decided to develop a skill where AI creates a draft based on existing manuals and other materials, and department promoters can proceed by reviewing and revising the content.

The approach is to look at the proposals and hypotheses organized by AI and think about "what is the actual situation" and "what would we like to change." Rather than thinking about what to write, the thinking shifts to how to revise this proposal to reflect the actual operations, which we believed would make it easier to proceed.

4. Details of the AsIs Skill

The first skill we worked on was the AsIs skill.

In creating the skill, we incorporated knowledge gained by BGU through past operational improvement support, as well as methods for creating AsIs and ToBe that we learned from the Business Transformation Division at Classmethod.
In working sessions, we learned about methods for organizing operations and expressing flows using materials actually used in operational improvements. Referencing these, we considered the creation procedures and output formats for the skill, reflecting them in aspects such as shapes, annotations, and page composition of the business flow.

We also adopted an approach where each department's manuals and lists of operations are used as input materials, AI organizes the draft, and department promoters review and revise it.

Classmethod Operations is primarily a remote work company, and materials explaining current operations — such as manuals, procedure documents, and operation lists — are available as data or online. From such materials, AI organizes operational information and presents hypotheses and confirmation items for areas that are lacking.

Department promoters respond to and revise the AI output.

Confirming What Cannot Be Understood from Manuals Alone

However, manuals and lists of operations do not necessarily represent the entire flow of operations.

For example, even if screen operations are described in detail, the operations handled by other departments, approval conditions, and responses to exceptions may be omitted. Also, manuals may be outdated and differ from current practices.

Therefore, in developing the skill, we emphasized that AI output should not be treated as finalized information as-is, and that people must review and approve the content.

The following is an explanatory example using a fictional application reception process.

Suppose a manual states: "Open email," "Confirm application content," "Open management ledger," "Enter required items and save," and "Send email to approver."

If each of these is placed in a separate box, the result is a diagram where operational steps are listed in sequence. With the skill, we consolidate these into units of work while confirming the responsible parties, decisions, and information inputs and outputs.

However, this organization alone does not reveal how to handle cases where there are deficiencies in the application, or what processing occurs after approval. AI therefore organizes confirmation items, and the department promoter verifies the content with the operations staff.

For example, points such as "Who is the application returned to if there are deficiencies?" "Is another review conducted after correction?" and "What constitutes completion of reception?" Based on the answers, tasks and branches are added or revised.

An example of a swimlane output from Cowork
(The skill outputs two types: an overall business flow + swimlane)

The above is an example of swimlane output.
Department promoters look at the generated proposals and confirm with operations staff whether "this task actually exists," "whether the responsible parties are correct," and "what happens under this condition." Whether the operations are correctly represented is judged by people.

Content directly written in the materials and content inferred by AI are also distinguished and recorded separately. This ensures that information not yet confirmed by the actual staff is not treated as verified fact.

Confirming Content in Text Before Creating the Diagram

In the AsIs skill, rather than drawing a swimlane diagram from the start, we structured the process so that operational content is first organized in Markdown text format, and that content is then reviewed.
This is because, while swimlane diagrams — which divide areas by person or department to show who performs which tasks — are easy to understand, we felt that once something is rendered as a diagram, it becomes harder to grasp "why that flow exists."
There was also the intent to reduce token usage by organizing the information to be exchanged, and to enable the organized text content to be readily understood by AI and utilized in the ToBe skill as well.
Therefore, the process was structured so that after the department promoter reviews and approves the text-organized content, it is then converted into a flow diagram in PowerPoint or similar tools.

It should be noted that there are still many challenges remaining with the generated diagrams. This point is explained in a later section.

5. Details of the ToBe Skill

In creating the ToBe skill, we first considered reviewing operations in line with the ECRS principles, rather than deciding upfront to replace with AI or automation.

It may seem inconsistent with the goal of AI Transformation to not presuppose AI adoption. However, we believe that AI and automation are means toward making operations better.

If a task itself is unnecessary, it is more appropriate to eliminate it rather than continue processing it with AI or automation. Introducing new mechanisms also generates additional maintenance and management tasks. The more AI and systems are used, the more costs are incurred.

Therefore, the improvement proposals generated by AI also follow a sequence where tasks are first reviewed, and then the roles of AI and automation are considered.

Consider Role Allocation After Reviewing Tasks

Improvement proposals are examined in the following order, in line with the ECRS concept.

Rather than replacing all current tasks with AI while leaving them intact, we first consider whether each task is necessary. Then we examine how to divide the remaining tasks among people, AI, and conventional automation.

Comparing a Realistic Proposal and an Ideal Proposal

In the skill, we structured it so that in principle two proposals are created: a "realistic, phased introduction proposal" and a "maximum AI utilization, ideal proposal."

This is to compare a proposal that is easy to implement with the current structure and systems against a proposal that uses AI agents and similar tools to reduce human input, transcription, and confirmation as much as possible.

For the earlier fictional application reception process, the following directions could be considered.

These are candidate explanations for illustrative purposes. Actual proposals are created after confirming the facts and constraints of the target operations.

We do not recommend the ideal proposal simply because it uses more AI. We compare expected effects along with costs, implementation period, coordination with other departments, and operational risks.

Rather than selecting only one, we also anticipate using the ideal proposal as a future goal and advancing in stages from the realistic proposal.

AI outputs two ToBe proposals from the AsIs flow. The intent is to confirm with the department and relevant parties whether the AI proposals are feasible.

Create Materials for Stakeholders to Make Decisions Before the Final Flow

In the ToBe skill as well, examination content is first organized as text. Furthermore, before creating the final ToBe flow, a PowerPoint document is created for comparing proposals and having stakeholders review them.

This is because department promoters alone cannot finalize the ToBe.
Coordination within the department and with related departments becomes necessary for changes to responsibilities, operational rules, and so on.
Swimlane diagrams are useful for understanding the flow of operations.
On the other hand, for multiple people to judge improvement proposals, it is also necessary to confirm the "purpose," "current issues," "reasons for change," and "conditions required for implementation."

Therefore, we structured the process so that materials confirming these points are created, a direction is selected, and then the detailed ToBe flow diagram is created. Where necessary, the technical implementation methods are also confirmed with the relevant IT departments and others.

Confirming content within the skill and proceeding to the next step is distinguished from organizational approval. For both AsIs and ToBe, we emphasize that people make the judgment of "this content is acceptable."

6. Skill Challenges and Future Improvement Goals

At this point, the output of the two skills is provided as a draft for department promoters to review and revise.

The positioning as a draft carries two meanings.

One is that department promoters and operations staff need to verify the operational content and improvement proposals organized by AI. This verification, I believe, will continue to be necessary throughout the creation process.

The other is that there are challenges with the flow diagram output — such as branching lines overlapping or text overflowing from boxes — requiring manual correction by people. The part of expressing operations containing complex conditions in a readable format using shapes and connecting lines is not yet sufficient.

Even when content is organized in Markdown, the task of verifying whether that content is properly reflected in the diagram, and revising the diagram as needed, still remains.

As shown above, output as-is has layout issues, and color is added after revision to make it easier to understand.

Reflecting Dialogue-Based Revisions in the Diagram

Going forward, we would like to improve the skill so that even for operations containing complex conditions, content confirmed and revised through dialogue can be properly reflected in the flow diagram.

The goal is a form where, if a user communicates a revision to a task or branch, the shapes and connecting lines are also updated accordingly, reducing the burden of manually correcting each element in PowerPoint.

To that end, we will use examples of operations containing complex branches and returns to verify whether the organized specifications and generated diagrams match, and whether stakeholders can follow the flow.

We will also revisit the content and manner of questions. Rather than requesting a large amount of information from the start, we would like to present the parts that can be organized from materials at hand, and allow for confirming missing points in sequence.

7. For Each Department to Advance Operational Reform

BGU's first initiative was to make it easier for department promoters in each department to start creating AsIs and ToBe.

AI organizes a draft from existing materials, and department promoters confirm and revise it with operations staff. Improvement proposals are also developed by having stakeholders select a direction while comparing candidates. To support this approach, we are developing two Cowork skills.

Currently, there are challenges with diagram output, and the output is not yet at a stage where it can be used directly as finished materials.
Going forward, while upholding the principle that people perform the confirmation of operational content and examination of improvement proposals, we will continue to improve the skill so that formatting and redrawing of materials can be handled by the skill as much as possible.

About Classmethod Operations Co., Ltd.

We are an operations company within the Classmethod Group.

We are a team of experts who, through "mechanisms" making full use of IT and AI, provide specialized teams in operations, maintenance development, support, information systems, and back office — offering everything from business process outsourcing to problem-solving and high-value-added services for our clients.

We are recruiting members for various positions.

If you are interested in the culture, mechanisms, and work style that together realize "Operational Excellence" and "Working and living in our own way," please visit the Classmethod Operations Co., Ltd. Corporate Site. ※ The company name was changed from Annotation Co., Ltd. in January 2026.