Have you ever considered the difference between a tool and a system?
Tools are things we use when we work. Systems, similarly, make work more efficient.
Some might have the image that a system is simply a more complex tool.
However, if we classify work into two types—iterative work and flow-based work—the distinction between a tool and a system becomes remarkably clear.
Iteration and Flow
Iterative work is the process of gradually creating a deliverable through trial and error, adapting flexibly as you go.
For iterative work, a toolkit that allows you to choose the right tool for specific tasks is useful.
Flow-based work, on the other hand, involves progressing through stages, producing the deliverable at the final stage.
For flow-based work, having a system to guide the work along the flow significantly improves productivity and quality.
Flow-Based Work Transformation and Systematization
Much of the work performed by humans is either iterative work or a component of a systematized flow-based process.
Transforming iterative work into flow-based work, and then systematizing it, significantly contributes to improvements in productivity and quality.
The Industrial Revolution and the IT Revolution
The Industrial Revolution and the IT Revolution are prime examples of significantly increased productivity and quality through the transformation of iterative work into flow-based work and its subsequent systematization.
Prior to the Industrial Revolution, manufacturing was performed as iterative work, where humans skillfully used tools, freely altering arrangements and procedures each time.
Information processing before the IT Revolution was also iterative work, with humans using tools and proceeding in a non-standardized manner.
By systematizing these processes, much like factory production lines and business IT systems, productivity and quality were enhanced.
However, not just systematization, but the flow-based transformation of that iterative work is extremely crucial. It was precisely because flow-based transformation was achieved that systematization became possible.
The Generative AI Revolution
When aiming to improve productivity and quality by utilizing generative AI in business, simply using AI as a tool will not yield true value.
The main objective is the transformation of iterative work into flow-based work, and then the systematization of that flow-based work.
Generative AI, capable of flexible adaptation, can handle iterative tasks. However, whether performed by humans or generative AI, there are limits to the productivity and quality of iterative work.
This is why it's crucial to aim for flow-based transformation and systematization.
One might argue that if flow-based transformation could improve productivity and quality even with human workers, such initiatives could have been undertaken before the advent of generative AI.
However, flow-based transformation premised on human workers is actually a very difficult problem. Human workers cannot immediately adapt to changes in task assignments or content.
On the other hand, when the worker is generative AI, it is easy to reconfigure assignments and task content through trial and error.
Unlike humans, generative AI can forget previous steps, instantaneously read and understand new procedures, and work based on them.
Therefore, the mainstream approach for leveraging generative AI in business will be the transformation of iterative work into flow-based work and its subsequent systematization.
Business Efficiency Improvement Using Generative AI
Let's consider an example of business efficiency improvement using generative AI.
As an example, consider the task of responding to employee inquiries about company rules.
By using generative AI, one can search company rules and draft answers.
However, there's a possibility that the generative AI might reference outdated rules or mistakenly imagine and provide answers not explicitly stated in the rules.
Moreover, inquiries come in various forms, such as email, messenger tools, phone calls, or verbal communication.
Therefore, employees handling inquiries still need to receive them as before.
It's conceivable that efficiency could be improved by answering questions that can be addressed on the spot, and for those requiring rule verification, inputting the inquiry content into generative AI to generate draft answers.
Additionally, for frequently asked questions, it's necessary to post them on the company's internal homepage as FAQs.
Generative AI can also be used to input typical questions and answers and create bulleted drafts for website publication.
Furthermore, when rule revisions are needed, generative AI can be utilized for drafting proposals.
Such applications might make a certain percentage of inquiry handling tasks more efficient.
However, this merely leaves the inquiry handling as iterative work and uses generative AI as a tool.
Consequently, the efficiency gains are very limited.
Flow-Based Work Transformation
To maximize the efficiency of the inquiry handling task given as an example, this task must be transformed into a flow.
To do this, the tasks performed by the person in charge when handling inquiries need to be detailed and formalized.
- Receive inquiries through various channels.
- If the inquiry is the same as one previously answered and there are no changes to related rules, provide the same answer.
- For new inquiries or inquiries involving rule changes, confirm the rules and draft an answer.
- Check that the draft answer does not refer to outdated rules or include information not stated in the rules.
- Check if approval is required before answering, and obtain approval if necessary.
- Respond via the channel through which the inquiry was received.
- Register the inquiry content, approval result, and answer result in the inquiry history data.
- Regularly check the inquiry history data and create drafts for updating frequently asked questions and answers.
- Update the internal company homepage after obtaining approval.
- Update the referenced rule data when rules are updated.
- Concurrently, record in the past inquiry history data that related answers and rule updates have occurred.
- Confirm if frequently asked questions and answers need review due to rule changes, and update if necessary.
By clearly defining the details of the tasks performed, as described above, these tasks can be connected, transforming flexible iterative work into a clearer flow-based process.
Example of Systematization
By creating this work-flow, the path to systematization becomes clear.
For systematization, if sacrificing some employee convenience is acceptable, one option is to consolidate inquiry channels.
Conversely, if employee convenience is prioritized, the system should maintain the ability to receive inquiries through all channels.
Basically, the system should directly receive inquiries. Only for verbal inquiries should the person in charge input them into the system.
After an inquiry is received, the IT system and generative AI should execute as much of the subsequent work as possible, following the flow. Initially, human checks and approvals should be interspersed throughout the system, and human operators should be able to make corrections.
Then, as the system is used for inquiry handling, if generative AI makes a mistake, instructions to the generative AI should be updated with warnings, points to check, examples of mistakes, and correct examples to prevent recurrence.
This allows for a reduction in generative AI errors. This process of updating instructions for generative AI can be made even more efficient if it is transformed into a flow-based task rather than an iterative one.
In this way, by systematizing flow-based work, even tasks that seemingly require human intervention can be replaced by a system centered around generative AI.
Common Misconceptions
Many people hold the view that generative AI's business application is not very effective at present, or that it's premature.
However, a significant number of these individuals often fall into two patterns of misunderstanding.
The first misconception arises from focusing on using generative AI as a tool.
As demonstrated here, leveraging generative AI as a tool for iterative tasks does not significantly boost business efficiency. Experiencing or hearing about this leads to this misconception.
The second misconception stems from focusing on having generative AI execute iterative tasks.
Indeed, trying to make current generative AI perform iterative tasks does not work well. Consequently, generative AI cannot fully take over the duties performed by humans, and focusing solely on this point leads to the misunderstanding.
Finally
As discussed here, by transforming iterative work into flow-based work and systematizing it, greater efficiency than with mere tools can be expected.
Furthermore, even if iterative work itself cannot be fully handled, many individual tasks within a flow-based process can be managed by current generative AI. Even if there are many initial errors, continuous improvement can be achieved by updating instructions.
Alternatively, tasks can be split as needed, separating drafting from checking, or implementing multi-stage checking.
If systematization can be achieved in this manner, then improvements will progress with each task, and operations will become more efficient over time.
This is a way of working that enables continuous improvement of the mechanism itself, similar to factory production and IT system implementation.
To leverage generative AI, a shift in mindset is required: instead of just improving your own iterative tasks, you need to objectively transform your work into flow-based processes and systematize them.