AIYING FOR OUTCOMES
Unit 01 — Start With the Outcome
TIME: 15–20 minutes | LEVEL: Start | NEED: Any conversational AI | MODE: Standalone
1. OUTCOME FIRST
What will you accomplish?
By the end of this unit, you will be able to turn something you need to do into a clear outcome before asking AI for help.
The principle:
Don't start with AI. Start with the outcome.
2. PRACTICE FIRST
Do before you read
Don't open AI yet.
Think of one real thing you need to do today.
For example:
Prepare a presentation.
Before asking AI to make the presentation, answer this yourself:
What should be different when I am finished?
Write your answer.
Perhaps:
I want the management team to understand our three priorities and agree on the next actions.
Now compare:
Task: Prepare a presentation.
Outcome: Management understands the priorities and agrees on next actions.
One describes what you will do.
The other describes what you want to achieve.
Predict
Before continuing, answer:
How might defining the outcome first change what you ask AI to do?
Keep your answer. You'll return to it later.
3. DISCOVER
What just happened?
A common way of starting with generative AI is:
What can AI do for me?
or:
What prompt should I use?
That produces a familiar sequence:
AI → PROMPT → OUTPUT
AIYING reverses the starting point:
OUTCOME → AI
First determine where you want to go.
Then decide how AI can help you get there.
Consider:
“Create a ten-slide presentation about our strategy.”
AI can probably produce ten competent slides.
But ten slides were never the real goal.
Compare it with:
“I want the leadership team to understand our three strategic priorities and leave the meeting agreeing on who owns each next action.”
Now you can ask much better questions:
What does the audience need to understand?
What information is essential?
What might confuse them?
What evidence would make the argument credible?
What should the final call to action be?
The outcome begins shaping the AI interaction.
Task, Output and Outcome are different
Task
Prepare a presentation.
↓
AI Output
Ten slides.
↓
Outcome
People understand the priorities and agree on next actions.
The distinction becomes even clearer elsewhere.
Learning
Task: Study KNN.
Output: AI explanation of KNN.
Outcome: I can explain KNN and correctly apply it to a simple classification problem.
Work
Task: Analyse customer complaints.
Output: AI-generated summary.
Outcome: Identify the recurring causes of complaints and decide which problem to address first.
Life
Task: Compare laptops.
Output: Comparison table.
Outcome: Choose a laptop that best fits what I actually need within my constraints.
This leads to one of the foundational ideas of AIYING:
AI can generate an output. That does not mean you have achieved an outcome.
4. POWER PROMPT
Clarify My Outcome
Use this before asking AI to solve an important task.
I need to [DESCRIBE THE TASK].
Before helping me do it, help me clarify what I am actually trying to accomplish.
Ask me focused questions, one at a time, about:
my desired outcome,
who or what is affected,
important context,
constraints,
what success would look like.
Do not solve the task yet.
Once you understand my situation, state my intended outcome in one clear sentence.
Then ask me whether that accurately represents what I want to accomplish.
Why “Do not solve the task yet”?
Because AI is extraordinarily good at starting production.
Sometimes the intelligent thing to do is prevent it from starting too soon.
Clarify first. Generate later.
5. PRACTICE EXERCISE
Now make it yours
Choose one real situation.
LEARN
Something you need to learn.
Instead of:
Explain neural networks.
Try:
I need to understand neural networks. Before teaching me, help me define what I should actually be able to understand or do when I have learned them.
WORK
Something you need to accomplish professionally.
Instead of:
Write a report on declining sales.
Try:
I need to analyse declining sales. Before preparing anything, help me clarify what decision or action this analysis should enable.
LIFE
Something you need to decide or accomplish personally.
Instead of:
Compare these three laptops.
Try:
I need to choose a laptop. Before comparing models, help me clarify what I actually need the laptop to accomplish for me.
Now use the Power Prompt.
When AI has clarified your outcome, write down:
BEFORE AI:
My task was ______________________________.
AFTER CLARIFICATION:
My outcome is ____________________________.
Now return to the prediction you made in Section 2.
What changed?
Did defining the outcome alter the way you would use AI?
Explain the difference without asking AI to answer for you.
6. JUDGE
Don't accept. Examine.
AI has proposed an outcome.
Don't automatically accept it.
Ask yourself:
Did AI understand what I really want?
What important context did I provide?
What context does AI still not know?
Did AI introduce assumptions?
Is the proposed outcome measurable or observable enough?
Is this actually my outcome—or has AI subtly chosen an outcome for me?
Now rate your confidence:
How confident are you that the outcome is correctly defined?
1 — Very uncertain | 2 | 3 | 4 | 5 — Very confident
If your confidence is below 4, continue clarifying.
This illustrates an important boundary:
AI can help clarify what matters.
Humans remain responsible for deciding what matters.
7. ACCOMPLISH
Turn the outcome into action
Now—and only now—ask AI:
“Given this agreed outcome, help me determine the best way AI can assist me in achieving it. Separate what AI can help with from what requires my own knowledge, judgment or action.”
Work on the task.
But keep your outcome visible.
At any point when the conversation starts expanding or drifting, ask:
Does what we are doing now directly contribute to the outcome we defined?
When you're finished, don't evaluate the session only by asking:
Was the AI output good?
Ask:
Did I accomplish the outcome?
If not, identify the gap:
Outcome → Current Result → Gap → Next Action
That next action begins another AIYING cycle.
8. REFERENCES & EVIDENCE
What's behind this idea?
Established evidence — Goal setting
Locke and Latham's extensive research on goal-setting theory shows that goals influence performance by directing attention and effort, increasing persistence, and encouraging the development or use of task-relevant strategies. Their review also emphasizes the importance of factors such as goal specificity, difficulty, feedback and commitment.
Established evidence — Human-centred AI
UNESCO's AI competency framework emphasizes a human-centred mindset alongside ethics, AI techniques/applications and AI system design. It explicitly places human agency, critical judgment and responsible use within AI competency rather than treating technical operation alone as sufficient.
Related thinking — End-User Computing
The history of End-User Computing demonstrates the importance of putting computational capabilities directly into the hands of people who understand the domain problem. It also produced important lessons concerning user education, support and appropriate controls.
AIYING proposition
The specific formulations:
Don't start with AI. Start with the outcome.
and
OUTCOME → THINK → ASK → JUDGE → ACCOMPLISH
are propositions of the AIYING for Outcomes methodology.
The cited research provides relevant intellectual foundations; it should not be interpreted as direct empirical validation of the AIYING framework itself.
9. FURTHER STEPS
Where can you go from here?
TRY
Look at three AI conversations you've had recently.
For each, identify:
Task → AI Output → Intended Outcome
Was the intended outcome ever explicitly stated?
EXPLORE
Start noticing the difference between:
Activity — what you are doing.
Task — what needs to be completed.
Output — what AI produces.
Outcome — what should change as a result.
READ
Explore established research on goal-setting theory, human-centred AI and End-User Computing. Pay particular attention to the distinction between giving people access to powerful technology and developing the capability to use it effectively.
NEXT
Unit 02 — Task ≠ Outcome
Explore one of the easiest mistakes to make with AI: asking it to complete the visible task without first discovering the outcome behind the task.
No comments:
Post a Comment