What if the problem isn’t AI’s answer, but the question we ask?
Generative AI can organize enormous amounts of information, analyze data, and produce convincing answers in seconds.
But even a sophisticated AI system cannot reliably help us solve the right problem if we begin with the wrong question.
In the AI era, writing effective prompts is not enough.
We need to determine where to focus our time and energy, what we really want to understand, which assumptions we are making, what evidence we can trust, and what decisions need to be made.
These are the capabilities supported by PQA (Precise Questioning and Answering).
In this article, Peemo, who researches publicly available information, and Dash, who analyzes leadership development, explore how PQA’s seven categories of questions can help us use AI more effectively and make better decisions.
1. Why Is the Ability to Ask Questions So Important in the AI Era?
Peemo (Research):
Generative AI has dramatically reduced the time required to gather information and conduct analysis.
However, asking AI a question does not automatically produce a useful answer.
Dash (Leadership Analysis):
Imagine a sales manager asking AI:
“Suggest ways to reduce the time salespeople spend preparing reports by 50%.”
AI might recommend automating data entry, improving reporting templates, or using voice recognition.
These may be useful suggestions.
But what if the manager asks a different question?
“Why do we need these reports in the first place?”
Now consider the same issue through PQA’s first category, Focus:
“Should our salespeople devote more of their limited time and energy to preparing reports or engaging with customers?”
This question shifts attention from improving the efficiency of reporting to reconsidering the priorities of the entire sales process.
Perhaps the reports are simply a legacy of an earlier management system and are rarely used in current decision-making.
If so, eliminating or redesigning the reporting process may create more value than making it faster.
AI generates answers to the questions we provide.
But deciding where to direct our attention and what questions deserve to be asked remains a critical human responsibility.
2. What Is PQA?
Peemo:
PQA is an approach to improving communication efficiency and critical thinking through precise questions and precise answers.
Its distinctive feature is the organization of questions into seven categories.
1. Focus: Determine where to direct time, energy, and attention.
2. Clarification:What exactly does this mean
3. Assumptions: What are we taking for granted?
4. Evidence: How do we know this is true?
5. Causes: What is producing this result?
6. Effects: What consequences or broader impacts might follow?
7. Action: Who should do what, and by when?
Dash:
These seven categories are not a checklist that must always be followed in sequence.
They provide a framework for selecting the questions most relevant to a particular situation.
The first category, Focus, is especially important.
We all work with limited time and energy.
As AI makes it possible to generate more information and options than ever before, deciding what deserves our attention becomes increasingly valuable.
For example, when an AI-generated recommendation seems questionable, we might immediately ask:
“Can you suggest another solution?”
But before doing so, we should consider:
“Are we working on the most important problem right now?”
“How much time and energy should we devote to this issue?”
“Is there another problem that deserves greater attention?”
We can then examine assumptions, evidence, and possible causes as needed.
The value of PQA is not in asking more questions.
It is in choosing the questions that matter and exploring them to the depth required for a sound decision.
3. Application 1: Using AI to Analyze Manufacturing Quality Problems
Peemo:
Imagine that a manufacturing department has experienced an increase in product defects.
An engineer asks AI:
“Recommend measures to reduce our defect rate.”
AI suggests inspecting equipment, reviewing operating procedures, and increasing inspection frequency.
Dash:
These are plausible actions.
However, implementing them before understanding the causes may produce limited results.
Let’s reconstruct the inquiry using PQA.
Focus:
“Among the quality problems we face, which one deserves the greatest concentration of time and attention right now?”
Clarification:
“Which products, processes, and time periods are affected by the increase in defects?”
Assumptions:
“Why are we assuming that equipment is responsible?”
Evidence:
“Do we have sufficient data to confirm the increase? Has the measurement method changed?”
Causes:
“What changes in materials, equipment, or operating conditions occurred when defects began to increase?”
Effects:
“If the problem continues, how will it affect delivery schedules, costs, and customer trust?”
Action:
“Which hypothesis should we test first? Who will verify it, and by when?”
These questions transform the interaction with AI from generating a list of countermeasures into a more disciplined investigation of priorities and possible causes.
A more effective AI prompt might be:
“Analyze the following manufacturing quality problem. First, identify which issues should receive priority attention. Do not assume the root cause is already known. Organize the facts that need verification, the underlying assumptions, and possible causal hypotheses. For each hypothesis, identify the data required to test it.”
The engineer must then compare AI’s hypotheses with actual production data and observations.
AI’s output should be treated as a set of hypotheses to investigate, not as a final conclusion.
That is one of the practical benefits of applying PQA.
4. Application 2: Discovering Customers’ Real Problems in Sales
Peemo:
Imagine a salesperson whose customer says:
“Your price is too high. We need a discount.”
The salesperson asks AI:
“How can I win this deal without reducing the price?”
AI recommends emphasizing product benefits and differentiating the offering from competitors.
Dash:
But this question contains an assumption.
It assumes that the customer’s primary problem is the purchase price itself.
PQA helps us challenge that assumption.
Focus:
“Should we concentrate our time and attention on price negotiations or on understanding the customer’s business challenges?”
Clarification:
“When the customer says the price is high, what are they comparing it with?”
Assumptions:
“Are we assuming that a discount is the only outcome the customer wants?”
Causes:
“Why is the customer requesting a price reduction now?”
Effects:
“Are there other problems affecting the customer’s business more significantly than the purchase price?”
These questions may reveal a different situation.
For example, the customer’s real problem might be production losses caused by unexpected equipment downtime.
In that case, improving maintenance services or delivery reliability could create greater value than reducing the price.
AI can suggest negotiation techniques.
But discovering the customer’s actual problem requires the salesperson to ask thoughtful questions and engage in meaningful dialogue.
5. Application 3: Improving the Impact of Leadership Development
Peemo:
Consider an HR department that receives high satisfaction ratings for its leadership training programs.
Despite the positive feedback, managers show little observable change in their workplace behavior.
An HR professional asks AI:
“How can we further improve satisfaction with our leadership training?”
Dash:
If satisfaction is already high, improving it may not be the most important issue.
Let’s use PQA to reconsider the problem.
Focus:
“Should we concentrate our time and energy on increasing training satisfaction or on encouraging behavioral change in the workplace?”
Clarification:
“What specific workplace behaviors do we expect participants to demonstrate?”
Assumptions:
“Are we assuming that attending training automatically leads to behavioral change?”
Evidence:
“What evidence supports the conclusion that behavior has not changed?”
Causes:
“Is the lack of change caused by insufficient skills, limited support from supervisors, or the workplace environment?”
Action:
“Who should provide follow-up support after training, and what should that support involve?”
These questions can reveal problems that cannot be solved by redesigning the training content alone.
For example, participants may learn effective feedback techniques but find few opportunities to practice them.
Their supervisors may not reinforce the desired behaviors, or everyday workload pressures may prevent them from applying what they learned.
In that situation, introducing supervisor follow-up and peer coaching may be more valuable than further increasing participant satisfaction.
Before asking AI to design a better training program, HR needs to determine what should actually change.
This is another important application of PQA.
6. Application 4: Evaluating AI Recommendations in Executive Meetings
Peemo:
Executive teams are increasingly likely to use AI to support market analysis and investment planning.
Imagine that AI recommends:
“The company should increase its investment in this market over the next three years.”
Dash:
An executive team cannot make a sound investment decision based solely on a convincing recommendation.
Let’s examine it using the seven PQA categories.
Focus:
“Should we concentrate our limited management resources and executive attention on this investment decision now, or are there more important priorities?”
Clarification:**
“What exactly does increasing investment mean in terms of scale, timing, and resources?”
Assumptions:
“Are we assuming that the market will continue to grow?”
Evidence:
“What data supports the growth forecast?”
Causes:
“What factors are driving the projected market growth?”
Effects:
“If the investment fails, what would be the consequences for our finances and workforce allocation?”
Action:
“What additional information must be verified, by whom, and before what deadline?”
One further question is especially useful:
“Is there anything important we have not asked about?”
This question helps expose possible blind spots.
It encourages decision-makers to consider risks, assumptions, or relevant information that may be missing from the AI-generated analysis.
Of course, AI cannot be expected to identify all of its own errors or omissions.
For important decisions, verification against primary sources, reliable data, and relevant professional expertise remains essential.
7. How Is Questioning AI Different from Questioning People?
Peemo:
PQA is designed to help people examine reasoning more rigorously.
However, the more penetrating a question becomes, the more likely another person may feel challenged or interrogated.
Dash:
This distinction matters.
PQA is not primarily intended for relationship-building or open-ended brainstorming.
Its purpose is efficient and effective intellectual exchange.
Therefore, even when the underlying question is the same, we should adjust how we ask it depending on whether we are interacting with AI or another person.
For example, we might ask AI:
“What evidence supports your conclusion? List all the assumptions behind it.”
That direct style is appropriate for an AI system.
But asking an employee the same question in a sharp tone may make them defensive.
Instead, a manager could say:
“I’d like us to examine how reliable this conclusion is. Could you walk me through the evidence that led you to it?”
The intellectual content of the question remains the same.
What changes is the tone and pace of the conversation.
This is also where DialogueWORKS provides a valuable perspective.
DialogueWORKS emphasizes understanding another person’s mental model and recognizing that our own interpretations are not necessarily facts.
PQA helps sharpen the precision of our questions.
DialogueWORKS helps us explore differences in understanding while maintaining constructive communication.
Together, they can improve not only how we work with AI but also how we communicate and make decisions with others.
8. Precise Answers Are Just as Important as Precise Questions
Peemo:
PQA has another essential dimension: precise answers.
One of its central principles is to communicate the core answer first and provide supporting details as needed.
Dash:
Imagine an executive asking a project manager:
“Is our AI implementation project on schedule?”
If the manager begins with a lengthy explanation of the project’s history and activities, the executive may struggle to understand the current situation.
A precise answer begins with the essential information:
“We are currently two weeks behind schedule. The primary cause is a delay in data preparation. We will finalize the recovery plan by next week.”
Additional details can then be provided if necessary.
Precise answers should also communicate uncertainty appropriately.
For example:
“Based on the information currently available, we believe the original completion date is still achievable. However, integration testing with the external system is not yet complete, so we cannot confirm the date with certainty.”
The same principle applies to reports and explanations generated by AI.
Longer answers are not necessarily better answers.
Clear communication of conclusions, supporting evidence, assumptions, uncertainty, and required actions improves the quality of decision-making.
9. How Can HR Integrate PQA into Employee Development?
Peemo:
The seven PQA categories can be applied not only to AI training but also to leadership development and everyday business meetings.
Dash:
One practical approach would be a three-month development program.
Month 1: Improve the Precision of Questions
Participants bring examples of questions they regularly ask AI.
They examine the assumptions behind those questions and reconstruct them using the seven PQA categories.
Particular attention is given to Focus:
“Which issues deserve the greatest investment of time and energy?”
Participants then compare how AI’s responses change when the questions are reformulated.
Month 2: Evaluate AI-Generated Answers
Participants examine the conclusions provided by AI, including their supporting evidence, assumptions, possible causes, and consequences.
When AI’s answers differ from workplace realities, participants investigate why those differences occur.
They learn to distinguish verified facts from plausible but unconfirmed interpretations.
Month 3: Apply PQA to Workplace Decisions
Each participant selects an actual business challenge and uses PQA to define and analyze the problem.
They then use AI to explore alternatives, evaluate options, and develop a decision and action plan.
During peer-coaching sessions, participants review one another’s questions.
“Does this question address the problem we really need to solve?”
“Where should we concentrate our time and energy?”
“Are we overlooking any important assumptions?”
“What information do we need before making a decision?”
Through repeated practice, PQA can become more than a conceptual framework.
It can become a practical capability that employees use in everyday work.
10. In the AI Era, Asking Better Questions Matters as Much as Finding Answers
Peemo:
Generative AI has made it possible to obtain more answers, more quickly, than ever before.
Dash:
But faster access to answers does not automatically produce better decisions.
In fact, the easier it becomes to generate convincing answers, the more important the quality of our questions becomes.
Where should we invest our time and energy?
What exactly are we trying to understand?
What assumptions are we making?
How do we know the information is reliable?
What is causing the problem?
What consequences might follow?
And what should we do next?
PQA’s seven categories provide a systematic way to approach these questions.
Before asking AI for a better answer, we need to formulate a better question.
After receiving AI’s answer, we need to continue questioning.
And we must translate the resulting insights into responsible decisions and actions.
In the AI era, employee development should go beyond teaching people how to operate AI tools.
It must also strengthen their ability to ask precise, relevant, and penetrating questions.
Is your organization teaching employees not only how to use AI, but also how to decide what questions are worth asking?
*Editorial note: The applications to AI, workplace scenarios, and proposed three-month development program represent SmartWorks’ analysis and recommendations.*
Akira Chida (Dash)
SmartWorks