Request Co., Ltd. (Headquarters: Shinjuku-ku, Tokyo, Representative Director: Tomoyasu Kobatake), which provides organizational behavior science® to companies, has organized new survey results on the characteristics of the top 1% of people who achieve results in the AI era, based on behavioral observations of 338,000 individuals and 980 companies.

* The illustrated materials for these survey results can be downloaded from the bottom of this release.

What has become clear from this survey is that simply being able to operate generative AI in detail does not fully explain the differences in performance. Generative AI has made it easier to quickly generate 'How' – the means – such as text creation, summarization, comparison, document creation, illustration, and idea generation.

On the other hand, for jobs with no single correct answer, or where there isn't just one correct answer, AI outputs are unlikely to lead to results without 'What' – the purpose – and 'Why' – the background: what is the objective, why is it necessary, and whose state will be changed.

In the behavioral observations of 338,000 individuals and 980 companies, what began to be seen in the top 1% who move work forward was not proficiency in AI itself. It was the ability to read the underlying assumptions and causal relationships from facts gathered through interactions with the field and others, identify problems and values that have not yet been articulated, frame them as objectives, and translate them into language that can be given to AI.

This survey's results organize signs that the value of people working in the AI era is beginning to shift from the ability to perform tasks to the ability to transform facts gained from experience into objectives and backgrounds.

Key Survey Findings

This survey, through behavioral observation and analysis of 338,000 individuals and 980 companies, identified the following five common signs in the top 1%:

1 With generative AI, 'How' – the means – has become easier to generate quickly.

Text, documents, summaries, comparisons, illustrations, idea generation, etc., can now be created quickly and with a certain level of quality by AI.

2 The difference in results is beginning to shift to 'before inputting into AI,' not the context of AI use.

Before deciding what to ask AI, the purpose of the request, whose state will be changed, and which facts will serve as the background are what determine the results.

3 For jobs with no single correct answer or multiple correct answers, 'What' – the purpose – and 'Why' – the background – are necessary.

In jobs with multiple options, or where the question itself is not yet defined, humans need to create the criteria for selecting the means provided by AI.

4 The top 1% transform facts gained from experience into purpose and background, rather than using experience as is.

They translate what they saw in the field, what they discussed with others, their sense of unease, their judgments, failures, and reactions into language that can be given to AI.

5 What companies need is not just AI training, but 'necessary experience design.'

It is necessary to intentionally increase experiences within work that create the purpose and background to be given to AI.

Who are the 'Top 1%' in this Release?

The 'top 1%' in this release does not refer to those proficient in AI operation.

In the behavioral observations of 338,000 individuals and 980 companies, this refers to the group where actions to move work forward were particularly strongly observed, from the perspectives of fewer confirmation exchanges, less rework, advancing the judgment of others, giving shape to ambiguous tasks, and mobilizing the actions of those around them.

What was common to this group was not just the speed of their work. It was the action of observing what to look at, who to converse with, which facts to confirm, and what the objective should be before starting the task.

Background: Generative AI Has Made 'How' – the Means – Easier to Generate

Generative AI quickly creates 'How' – the means. Refining text. Making documents easier to understand. Summarizing information. Creating comparison tables. Organizing discussion points. Generating illustration ideas. Drafting email correspondence. Creating a draft proposal. These have become easier to execute faster and with a certain level of quality than before.

However, faster work does not equate to achieving results. If the time saved by AI is merely reinvested in current work for incremental improvements, it does not lead to work that creates new value. What AI generates first is 'work capacity.' For this capacity to become 'strategic capacity,' organizations need to redesign their objectives, roles, responsibilities, evaluations, and methods of mobilizing people.

This survey's results address the question that lies even further upstream:

To what purpose should that capacity be directed in the first place? From which facts does someone draw that purpose?

This is where new performance differences in the AI era are beginning to emerge.

Survey Result 1: Performance Differences Are Beginning to Emerge 'Before Inputting into AI'

Previously, performance differences were often apparent in how quickly and accurately documents could be created, how clearly text could be written, or how accurately reports could be summarized.

However, as generative AI makes 'How' easier to generate, performance differences are shifting to the following questions:

What should be created?

For whom should it be created?

To advance whose judgment should it be created?

What are the difficulties the other party has not yet articulated?

What facts, assumptions, or causal relationships lie in the background?

In the AI era, those who achieve results are not just those proficient in AI. They are people who can identify values not yet articulated and articulate them as purpose and background before asking AI to perform tasks. In other words, performance differences are beginning to emerge not after AI output, but in the 'ability to transform facts gained from experience into purpose and background' before inputting into AI.

Survey Result 2: Jobs with Correct Answers Speed Up with AI. Jobs Where People Create Performance Differences Will Shift to Jobs Without a Single Correct Answer or With Multiple Correct Answers.

Future work will broadly divide into three categories:

1 Jobs with Correct Answers

These are jobs where the purpose, method, criteria, and completion conditions are relatively clear. Examples include creating standard text, summarizing, creating comparison tables, organizing meeting minutes, adjusting formats, and processing according to existing rules.

In this domain, generative AI strongly supports 'How' – the means. The human role shifts from execution itself to confirmation, quality assurance, and final judgment.

2 Jobs with Multiple Correct Answers

These are jobs where the purpose or challenge is somewhat visible, but there are multiple options, and the choice depends on the situation. Examples include which customers to prioritize, which measures to start with, which document structure will resonate with the audience, and whether to prioritize short-term results or long-term value.

In this domain, AI can generate many proposals. However, the choice of which proposal to select depends on the purpose, background, constraints, and value criteria.

3 Jobs Without Correct Answers

These are jobs where the questions, purpose, problems, and value criteria are not yet clear. Examples include identifying problems that customers have not yet articulated, finding new themes from a sense of unease in the field, discovering values that have not yet become markets, and deciding which new purposes to direct the capacity generated by AI towards.

In this domain, asking AI for answers directly is insufficient. Before asking AI, humans need to observe facts, hypothesize backgrounds, and define the purpose.

Jobs where people will create performance differences in the future are beginning to shift towards jobs without correct answers and jobs with multiple correct answers. What will be asked is not proficiency in AI, but the ability to read facts gained from experience and create 'What' – the purpose – and 'Why' – the background.

Basic Actions Common to the Top 1%

'Experience → Facts → Background Hypothesis → Purpose Formation → Articulation → AI Utilization → Judgment Update'

The top 1% are beginning to exhibit the following basic actions:

1 Experiencing

Observing the field. Talking with others. Having a sense of unease. Making judgments. Observing results.

2 Extracting Facts

What happened? Who was having trouble? What was stalled? What was repeated?

3 Hypothesizing Background

Which assumptions were misaligned? What was causally linked to what? Why did this problem occur?

4 Forming Purpose

Whose state will be changed, and how? What is this work intended to advance?

5 Articulating and Giving to AI

Articulating the purpose, background, facts, assumptions, causal relationships, constraints, and expected changes.

6 AI Generates 'How'

Generating text, documents, illustrations, comparisons, options, procedures, proposals, diagnoses, design drafts.

7 Humans Judge and Return to the Field

Selecting, correcting, testing, and updating AI outputs in light of the purpose, background, and field facts.

This series of steps is beginning to become the basic workflow for work in the AI era.

What Are 'Values Not Yet Articulated'?

'Values not yet articulated' in this release refer to problems, uncertainties, anxieties, misaligned assumptions, lack of decision-making information, and desired states that the other party or the field has not yet fully verbalized.

For example, when a client says, 'I want the document to be easier to understand,' what may truly be needed is not just making the document look better, but providing decision-making material that is easy to explain to superiors and stakeholders.

When the field says, 'We're busy,' what may actually be happening is not just a high workload, but ambiguity about what to prioritize, how much autonomy they have in decision-making, and whom to consult about what.

Generative AI can provide means based on the information given. However, identifying values not yet articulated requires fact-finding in the field, dialogue with others, a sense of unease, and judgment experience.

Work Styles Are Shifting from 'Performing Tasks' to 'Creating Purpose'

In past organizations, the ability to execute correctly, quickly, and meticulously against given objectives was emphasized. Of course, execution ability will continue to be necessary.

However, as generative AI begins to handle 'How,' the central value of workers changes. Processing assigned tasks. Following established procedures. Creating requested documents. Refining emails and reports. These alone are unlikely to create performance differences in the AI era.

What will become important is the stage before work becomes work. Who is having trouble? What has not yet been articulated? Which assumptions are misaligned? What needs to be changed to advance the judgment of the other party or organization? For that, what should be delegated to AI?

Work styles are shifting from 'performing assigned tasks' to 'transforming unarticulated values into purpose.'

What Companies Need is Not Just AI Training, but 'Necessary Experience Design'

Introducing AI tools and prompt training are important. However, they are not sufficient on their own.

If experiences that create 'What' – the purpose – and 'Why' – the background – to be given to AI are lacking in work, even if the AI output is polished, it will be difficult to influence the judgment of others or the organization.

What companies need is to design work so that employees can gain the following experiences: go to see facts, converse with others, identify misaligned assumptions, hypothesize causal relationships, define purpose, make small judgments, update judgments based on results, and translate them into language that can be given to AI.

This is called necessary experience design. Necessary experience design is a concept that intentionally increases opportunities within work for dialogue based on relationships with others, confirmation of background facts, assumptions, and causal relationships, small judgments, updating judgment results, and articulation for AI, in order to foster individuals who can articulate the purpose and background to be given to AI.

What is needed in the AI era is not just training in AI operation. It is to intentionally increase experiences before giving input to AI within work.

This Survey's Results Can Be Organized into Three Perspectives for Understanding Performance Differences in the AI Era:

1. Performance differences in the AI era are born before inputting into AI.

Experience ➡ Facts ➡ Background Hypothesis ➡ Purpose Formation ➡ Articulation ➡ AI Utilization ➡ Judgment Update

The more AI generates 'How,' the more human performance differences emerge 'before inputting into AI.'

2. Generative AI handles 'How,' while humans create 'What' and 'Why.'

How: Text creation, summarization, comparison, document creation, illustration, idea generation

What: Whose state will be changed, and how?

Why: Why is this necessary now? Background: Fact-checking, dialogue, assumptions, causal relationships, judgment experience

3. Five Experiences Common to the Top 1%

1 Experience building relationships

2 Experience confirming background facts, assumptions, and causal relationships through dialogue

3 Experience creating purpose

4 Experience making small judgments and updating based on results

5 Experience articulating into language understandable by AI

AI Era Purpose Formation Capability Quick Check

In conjunction with this release, a 'Quick Check for Purpose Formation Capability in the AI Era' has been organized to allow companies and workers to assess their current standing.

Example Check Items:

Before requesting AI, confirming whose state needs to be changed and how.

Articulating background facts and assumptions, not just task instructions.

Having opportunities to listen to problems that the other party has not yet articulated.

Modifying AI-generated proposals into forms usable in the field.

Reflecting on judgment results and transforming them into the next question or judgment criteria.

Being able to provide AI with purpose, background, facts, constraints, and expected changes.

This quick check allows for an assessment of the level of purpose formation capability before inputting into AI, rather than AI utilization itself.

Illustrated Materials for These Survey Results Can Be Downloaded as a PDF

The 14 main illustrations introduced in this release have been compiled into A4 landscape format illustrated materials, suitable for easy internal sharing.

These materials allow for a one-by-one review of the structure of performance differences in the AI era, basic actions common to the top 1%, necessary experience design, and the purpose formation capability check, and can be utilized for management meetings, AI utilization promotion, human resource development strategy discussions, and internal study sessions.

d68315-202-2c2199658922b20ef3e1faa93c9e77c4.pdf

Representative's Comment

Generative AI has significantly changed the 'How' of work. Refining text, creating documents, comparing, illustrating, and generating ideas. These can now be done faster than before.

However, for jobs with no single correct answer or multiple correct answers, 'How' alone does not lead to results.

Whose state will be changed, and how?

Why is this necessary?

What facts lie in the background?

Which assumptions are misaligned?

What is causally linked to what?

These are not things that AI can automatically determine. They must be framed as purpose and articulated by humans based on facts gained in the field, dialogue with others, and accumulated judgment experience.

In the behavioral observations of 338,000 individuals and 980 companies, what was seen in the top 1% was not proficiency in AI itself, but the tendency to identify values not yet articulated before inputting into AI.

In the AI era, the value of people working will no longer be measured solely by work volume. Those who can transform the field they have observed, the voices they have heard, the unease they have felt, and the judgment experience they have accumulated into purpose and background will be the ones who best leverage AI.

Survey Overview

Survey Name: Survey on Behavioral Signs of People Achieving Results in the AI Era

Survey Sponsor: Request Co., Ltd. Judgment Design Laboratory

Data Subject: Behavioral observation data of workers accumulated by Request Co., Ltd. through corporate training, practical support, behavioral observation, reflection entries, practical assignments, and dialogue records.

Data Scale: 338,000 individuals, 980 companies

Analysis Scenes: Reporting, consultation, meetings, email/chat, request handling, customer service, trouble response, handling busy periods, AI utilization scenarios, etc.

Analysis Perspectives: Confirmation exchanges, rework, actions that advance the judgment of others, actions that give shape to ambiguous tasks, pre-task fact-checking, purpose formation, background articulation, updating judgment experience, etc.

Note: The 'top 1%' in this release does not refer to the ranking of AI operation skills, but to the group where actions to move work forward were particularly strongly observed in the behavioral observations of 338,000 individuals and 980 companies. This survey is not an evaluation of the performance of generative AI itself, but an organization of where performance differences among workers are beginning to emerge in an era where generative AI enters the workplace, from the perspective of behavioral observation.

How We Can Help

Based on these survey results, Request Co., Ltd. is providing the following support to companies:

Detailed Check for Purpose Formation Capability in the AI Era

Diagnosis of Necessary Experience Design in the AI Era

'Job Design for Creating Purpose and Background' Training for Managers

Practical Course 'Articulating Purpose and Background for AI Input' for Field Staff

Organizational Support to Connect Generative AI Utilization from Task Efficiency Improvement to New Value Creation

The next challenge in AI utilization is not just tool introduction or prompt writing. It is about what facts people observe, what backgrounds they infer, and what purpose they define before inputting into AI. It is about how to increase those experiences within work.

The more generative AI handles 'How,' the more human performance differences are shifting to the ability to read facts gained from experience, create 'What' – the purpose – and 'Why' – the background – and provide them to AI.

For inquiries regarding the Detailed Check for Purpose Formation Capability in the AI Era, Necessary Experience Design Diagnosis, training courses, and practical courses, please contact the Judgment Design Laboratory.

About Request Co., Ltd.

Request Co., Ltd., under the banner of 'Aiming for Better,' is a company with eight research institutes, based on organizational behavior science®, derived from the work experience data of 338,000 workers at 980 companies (as of June 2026).

We conduct research and educational development to elucidate 'why' the thoughts and actions of people working in organizations occur and persist, from their business environment, history, and experiences, and to better reproduce them.

Company Name: Request Co., Ltd.

URL: https://www.requestgroup.jp/

Representative Director: Tomoyasu Kobatake

Location: Keio Frento Shinjuku 3-chome 4F, 3-4-8 Shinjuku, Shinjuku-ku, Tokyo

E-mail: request@requestgroup.jp

FACT BOX

  • Source: PR TIMES
  • Category: Survey結果