Summary: This Intermediate Guide for the ONHT (Objective, Needs, How, Trajectory) Framework transforms you from someone who uses GenAI into someone who thinks with GenAI by adding the missing cognitive functions that current GenAI lacks. The framework works through three critical pillars – Empathy (understanding all stakeholders), Critical Thinking (challenging assumptions), and Human in the Loop (active partnership). Master these patterns and you'll be solving complex problems others can't even approach, becoming indispensable by designing interactions that produce exceptional results rather than just functional outputs.
This guide transforms you from someone who uses AI to someone who thinks with AI. The difference changes everything.
Why Intermediate Skills Matter
Here's what happens at most companies right now:
Beginners get functional outputs. Fine for simple tasks. Intermediates solve complex problems. They become indispensable. The gap? Understanding that AI responds to structured thinking patterns.
Master these patterns, and you'll handle challenges others can't even approach.
The Shift in Thinking
The Beginner mindset: "How do I get AI to do this task?"
The Intermediate mindset: "How do I design an interaction that produces exceptional results?"
It's a fundamental shift in approach. The results show why it matters
Let's build that cognitive prompt design capability.
The Missing Pieces: Why ONHT Works
Back in 2023, I recognised something crucial about AI's limitations. GenAI systems are missing key brain functions that humans use naturally. Understanding these gaps, and how to bridge them, is what makes ONHT so powerful.
Episodic Memory: Remembering specific interactions
Situation Models: Understanding context deeply
You shouldn't consider these as technical limitations. They're opportunities.
By recognising these five missing functions, you can design prompts that fill the gaps. That's the real power. When you provide the context, memory, self-awareness and reasoning that AI lacks, you'll get results that others can't even imagine.
Slide from my 'The future of AI in Enterprise' presentation at HumanMade's AI event
How ONHT + The 3 Pillars Bridge These Gaps
The ONHT framework, combined with three critical pillars, adds these missing functions back:
1. Empathy bridges the situation model gap by ensuring we provide rich human context 2. Critical Thinking adds meta-cognition through structured reasoning approaches 3. Human In The Loop (HITL) provides the self-monitoring and orchestration AI lacks
Throughout this guide, you'll see how each element addresses specific cognitive gaps. This isn't just about better prompts - it's about creating a complete thinking system.
The Three Pillars in Practice
Before diving into advanced ONHT, let's understand how these pillars transform your results:
Empathy: Understanding All Stakeholders
Not just "be nice to the AI." Real empathy means:
Understanding your audience's pressures and fears
Recognising the AI's need for context
Anticipating human reactions to AI output
Critical Thinking: Strategic Scepticism
Not accepting first answers. Instead:
Challenging assumptions in your prompts
Building validation into your trajectory
Questioning what's missing, not just what's there
Human In The Loop: Active Partnership
You're not just prompting. You're:
Guiding discovery through intelligent iteration
Catching nuances AI misses
Translating between AI logic and human needs
These pillars aren't add-ons. They're integral to making ONHT work at an intermediate level.
Objective: Three Levels of Depth
Objectives have layers. Peel them back.
Quick Example: Bad vs Good
Bad Objective: "Help me with our customer retention problem"
Good Objective: "Create a 90-day retention strategy to reduce our 23% enterprise churn rate, with week-by-week implementation plan and success metrics"
The good version provides clear scope, timeline, and success criteria. Now let's build sophistication through three levels.
Level 1: Surface Understanding
Example: "Help me improve our customer retention strategy"
This states a need but lacks depth. AI will give you generic advice about customer satisfaction.
Level 2: Strategic Context
Example: "Our SaaS platform lost 23% of enterprise customers last quarter. Board wants a retention strategy by Friday. They're considering cutting our department if we can't show improvement in 90 days."
Now we have urgency and stakes. AI understands this isn't an academic exercise.
Level 3: Complete Strategic Picture
Example:
[MULTI-LAYERED OBJECTIVE]
Surface: Create customer retention strategy
Business Reality: 23% enterprise churn threatens our survival
Political Context: Board meeting Friday, CEO under pressure from investors
Human Impact: 12-person team jobs at risk, morale already low
Hidden Agenda: CFO pushing to outsource customer success
Success Criteria:
- Show path to <10% churn in 90 days
- Prove ROI within current budget
- Get board champion (ideally the CMO who likes us)
My Role: VP Customer Success, 18 months tenure, hired to fix this
Resources: Team of 12, £200K quarterly budget, good product but poor onboarding
Level 3 reveals the complete chess board. AI can now think strategically, not just tactically.
See how the 3 pillars strengthen this:
Empathy: Understands all stakeholder pressures and fears (board, CEO, team morale)
Critical Thinking: Challenges surface assumptions to reveal real needs (it's not just about churn)
HITL: You've provided context AI would never request on its own
Needs: Information Architecture
Think of your needs section as information design. Not just what to include, how to structure it for human decision-making.
Quick 'Needs' Example: Bad vs Good
Bad Section for 'Needs': "We have high churn. Customers complain. Competitors are better. Budget is tight."
Good Section for 'Needs': "Financial: 23% churn = £2.3M ARR lost. Root cause: 67% never complete onboarding. Competition: 8% churn with 2-day setup vs our 2 weeks. Constraint: £200K budget, team of 12."
The good version structures information for immediate understanding and action. Let's develop this through three levels.
Level 1: Basic Information Dump
Level 1 'Needs' Example:
[NEEDS]
- Churn is 23%
- Customers complain about onboarding
- Competitors have better support
- We need to fix this
Scattered facts without structure. AI struggles to prioritise.
Level 2: Organised by Category
Level 2 'Needs' Example:
[NEEDS]
Customer Data:
- Q3 churn: 23% (up from 15% in Q2)
- Main complaints: onboarding (45%), support response (30%), missing features (25%)
Competitive Position:
- Main competitor: 8% churn rate
- They offer 24/7 support, we offer business hours only
- Their onboarding: 2 days, ours: 2 weeks
Better organisation, clearer patterns. AI can identify priorities.
Level 3: Strategically Architected Information
Level 3 'Needs' Example:
[NEEDS - Structured for Executive Decision-Making]
FINANCIAL IMPACT LAYER (what the CFO sees):
- Q3 churn: 23% = £2.3M ARR lost
- Customer lifetime value: £127K (was £198K last year)
- Cost to acquire: £15K (up from £11K)
- Current burn: £28K per lost customer
- Competitor comparison: They profit £43K per customer, we lose £28K
ROOT CAUSE LAYER (what really drives churn):
Week 1-2: Onboarding chaos
- 67% of churned customers never completed setup
- Average time to value: 21 days (industry: 7)
- Technical issues: 34% need engineering help
- Success metric: Only 23% reach "aha moment" in first month
Week 3-8: Support frustration
- Response time: 48-72 hours (promised: 4 hours)
- First-call resolution: 34% (was 67% with full team)
- Escalations: 3x increase since layoffs
- NPS during this period: -23 (yes, negative)
Month 3+: Feature gaps
- Mobile app: Promised Q1, delayed to Q4
- API limitations: Can't integrate with 3 major platforms
- Reporting: Manual exports only, no dashboards
- Competition: Has all these features at lower price
POLITICAL LAYER (what influences decisions):
- CFO: Wants to outsource (has preferred vendor)
- CMO: Ally, needs customer success for her metrics
- CEO: Gave us 90 days, reports to aggressive board
- Sales: Blaming us for missed targets, very vocal
- Product: Overwhelmed, can't commit to roadmap
OPPORTUNITY LAYER (where we can win):
- Quick wins: 5 customers worth £400K ARR salvageable this week
- Process fix: Onboarding can be cut to 5 days with automation
- Team morale: Will improve with clear plan and small wins
- Board perception: CMO will champion if we show week 1 progress
Level 3 structures information in the way executives think and make decisions. Each layer serves a purpose, building a complete picture. If you can adapt your output structure to a required and relevant format, this will improve its impact on the intended audience.
The 3 pillars in your 'Needs' section at work:
Empathy: Understands all stakeholder pressures and fears (board, CEO, team morale)
Critical Thinking: Challenges surface assumptions to reveal real needs (it's not just about churn)
Human In The Loop: You've provided context AI would never request on its own
How: Beyond Simple Thinking
Basic "how" tells AI what framework to use. Sophisticated "how" teaches AI exactly how to think within the ONHT framework.
Quick 'How' Example: Bad vs Good
Bad How Section: "Think like an expert"
Good How Section: "Think like a turnaround specialist who's saved 3 companies. Balance quick wins for political capital with sustainable fixes. Challenge every assumption, but stay constructive."
The good version provides a specific mindset, approach, and constraints. Now let's build depth.
Level 1 'How': Name the Approach
Level 1 'How' Example:
[HOW]
Think like a customer success expert
Too vague. AI defaults to generic expertise.
Level 2 'How': Define the Thinking Style
Level 2 'How' Example:
[HOW]
Think like a Silicon Valley customer success leader who's turned around failing departments. Focus on quick wins while building long-term systems. Balance empathy for team with urgency for results.
Better guidance on mindset and priorities. AI adopts a specific perspective.
Level 3 'How': Complete Thinking Architecture
Level 3 'How' Example:
[HOW - Cognitive Framework Stack]
PRIMARY LENS: Battle-tested VP Customer Success who's saved 3 companies
- Been through "90 days to prove it" pressure
- Knows how to navigate political landmines
- Balances quick wins with sustainable change
- Protects team while driving performance
ANALYSIS APPROACH:
1. Start with undeniable financial impact (CFO can't argue with maths)
2. Show customer voice data (CEO cares about market perception)
3. Present competitor benchmarks (board loves comparisons)
4. Map quick wins to key champions (CMO needs wins too)
CRITICAL THINKING FILTERS:
- Challenge: "What would a hostile CFO attack in this plan?"
- Validate: "Which customers would actually confirm our assumptions?"
- Stress-test: "What breaks if we lose 2 more team members?"
- Politics: "How does each stakeholder win from this?"
PRESENTATION MINDSET:
- Confidence without arrogance (we screwed up but can fix it)
- Data-driven but human (customers are people, not numbers)
- Urgent but not panicked (controlled intensity)
- Own the past, control the future
DECISION ARCHITECTURE:
- Every recommendation has 3 options (give them choice)
- Every option has clear trade-offs (no hidden surprises)
- Every trade-off has risk mitigation (show we've thought ahead)
- Every mitigation has an owner (accountability matters)
SUCCESS VISUALISATION:
Before proposing anything, imagine:
- CFO's objections and pre-empt them
- CEO's questions and answer first
- CMO's needs and align with them
- Team's fears and address them
Level 3 doesn't just name an approach, it builds a complete thinking system that addresses GenAI's meta-cognitive gaps.
Activating the 3 pillars in the 'How' section of your prompt:
Critical Thinking: Built-in challenge filters and decision architecture ("What would a hostile CFO attack?")
Human In The Loop (HITL): You're teaching (explaining to) AI how to think, not just what to think
Trajectory: Designed Discovery
Different problems need different journeys to be successful. The trajectory is where your three pillars: Empathy, Critical Thinking, and Human In The Loop come alive.
Quick 'Trajectory' Example: Bad vs Good
Bad Trajectory: "Analyse the problem and give me solutions"
Good Trajectory: "First, identify top 3 churn drivers from data. Then, design interventions for each. Prioritise by impact/effort. Create week-by-week implementation plan. Include early warning systems."
The good version guides discovery with clear phases and outputs. Let's develop sophistication.
Level 1 'Trajectory': Linear Exploration
Level 1 'Trajectory' Example:
[TRAJECTORY]
1. Analyse customer feedback
2. Identify main issues
3. Propose solutions
4. Create an action plan
Basic steps, no adaptation. Misses nuance.
Level 2 'Trajectory': Structured Discovery
Level 2 'Trajectory' Example:
[TRAJECTORY]
1. Scan all feedback for patterns (broad survey)
2. Deep dive into top 3 issues (focused analysis)
3. Compare with successful customers (contrast study)
4. Generate solutions for each pattern
5. Prioritise by impact and effort
6. Build phased implementation plan
More sophisticated, but still linear. Beginning to guide thinking.
Level 3 'Trajectory': Adaptive Discovery System
Level 3 'Trajectory' Example:
[TRAJECTORY - Dynamic Discovery Protocol]
PHASE 1: Rapid Pattern Recognition
1. "First, identify the 3 most expensive problems (financial lens)"
2. "Then find the 3 most emotional problems (customer lens)"
3. "Where do these overlap? That's our focus zone"
PHASE 2: Intelligent Deep Dive
IF mainly operational issues:
- Map process breakdowns
- Calculate automation potential
- Design surgical fixes
IF mainly product gaps:
- Prioritise by customer segment value
- Identify MVP solutions
- Create buy-vs-build analysis
IF mainly people/culture issues:
- Assess team capability gaps
- Design rapid training
- Plan cultural interventions
PHASE 3: Solution Architecture
Based on what you discovered above:
- Quick wins: What fixes this week? (for political capital)
- Foundation: What prevents recurrence? (for sustainability)
- Innovation: What leapfrogs competition? (for differentiation)
PHASE 4: Stakeholder Alignment Map
For each solution:
- CFO angle: ROI within 60 days?
- CEO angle: Market story improvement?
- CMO angle: Brand impact positive?
- Team angle: Morale boost likely?
PHASE 5: Risk-Adjusted Roadmap
Build implementation plan that:
- Assumes 30% less resources (CFO will cut)
- Plans for 2x timeline (things go wrong)
- Has kill switches (if not working)
- Shows progress weekly (maintains support)
VALIDATION GATES:
After each phase: "What would kill this plan?"
If answer exists, design around it before proceeding.
Level 3 trajectories need to be adapted based on your discoveries and you also need to build in your internal political/practical reality. This is where Human In The Loop shines: you're actively guiding the exploration.
The 3 pillars guide the 'Discovery' section:
HITL: Adaptive paths based on what you discover and your reality (IF/THEN logic throughout)
Critical Thinking: Validation gates prevent costly mistakes ("What would kill this plan?")
Empathy: Each phase considers human and political reality (stakeholder alignment map)
Real-World Pattern Library
Let's see complete ONHT patterns at all three levels (Level 1 to Level 3), showing how the framework and pillars work together.
The Stressed Executive Pattern
Bad Example: Help executives with digital transformation
Good Example: [OBJECTIVE] Create 30-day digital transformation wins for sceptical board [NEEDS] Budget: £500K, Team: 8 people, Timeline: 90 days max [HOW] Think like McKinsey partner - clear, confident, results-focused [TRAJECTORY] Week 1 wins ? 30-day foundation ? 60-day momentum ? 90-day transformation
Now let's build a more thorough version:
Level 1: Basic Transformation Strategy Response
[OBJECTIVE]
Help executives see quick wins in digital transformation
[NEEDS]
- Board wants results
- Budget is tight
- Team is small
[HOW]
Think like a consultant
[TRAJECTORY]
List some quick wins
[OBJECTIVE]
Create 30-60-90 day plan showing digital transformation progress to sceptical board
[NEEDS]
- Board patience: 90 days maximum
- Budget: £500K total (no more)
- Team: 8 people, mixed skills
- Current state: 2 failed initiatives last year
- Political climate: CEO's job on line
[HOW]
Think like McKinsey partner specialising in digital turnarounds. Focus on visible wins that build credibility.
[TRAJECTORY]
1. Week 1-2 wins: What can we show immediately?
2. 30-day foundation: What systems enable scale?
3. 60-day momentum: What proves the strategy?
4. 90-day transformation: What changes the game?
Much better structure and context. AI can develop actionable plan.
[OBJECTIVE - Multi-dimensional]
Surface need: 90-day digital transformation proof points
Real need: Save CEO's job and our department
Ultimate need: Position company for acquisition or IPO
Human reality: Team exhausted from failed projects, board hostile, competitors gaining
[NEEDS - Layered Intelligence]
CRISIS PARAMETERS:
- Day 0 reality: Stock down 30%, activist investor circling
- Day 90 target: Show path to 20% efficiency gain
- Budget reality: £500K approved, not a penny more
- Team reality: 8 people, 3 flight risks, 2 underperformers
POLITICAL DYNAMICS:
- CEO: Hired us, needs wins to survive vote
- CFO: Wants to cut costs, sees IT as expense
- Board Chair: Tech-savvy, could be ally if approached right
- Activist: Pushing for leadership change, watching everything
COMPETITIVE PRESSURE:
- Main rival: Just announced AI integration (stock up 15%)
- Market expects: Our response within 30 days
- Analysts asking: "Why so far behind?"
- Customers nervous: Two large ones "evaluating options"
SUCCESS INDICATORS:
- Week 1: Announce something that moves stock
- Day 30: Customer win based on new capability
- Day 60: Efficiency metrics trending right
- Day 90: Clear path to market leadership
[HOW - Transformation Leadership Framework]
Channel composite mindset:
- Wartime CEO (peacetime is over)
- McKinsey senior partner (strategic clarity)
- Startup founder (speed and scrappiness)
- Political operative (navigate stakeholders)
Decision filters:
- Speed over perfection
- Visible over important
- Allies over process
- Stories over statistics
Communication style:
- Confident but not cocky
- Urgent without panic
- Clear without oversimplifying
- Inspiring despite reality
[TRAJECTORY - Orchestrated Campaign]
WEEK 1: Shock and Awe
1. Identify one thing we can announce immediately
(Even if just accelerating existing plan)
2. Draft CEO email to company: "Digital First" vision
3. Create visual: Before/After state
4. Leak to friendly analyst: "Transformation underway"
DAY 8-30: Foundation Sprint
Build based on Week 1 reaction:
IF positive response:
- Double down on that direction
- Recruit board chair as sponsor
- Get customer testimonial
IF sceptical response:
- Pivot to efficiency angle
- Show cost savings first
- Get CFO buy-in
DAY 31-60: Momentum Machine
- Weekly wins announced internally
- Bi-weekly board updates showing progress
- Customer advisory board formed
- Competitor comparison showing gap closing
DAY 61-90: Victory Lap Setup
- Major customer win announcement
- Efficiency gains documented
- Team expansion justified
- Next phase funding requested
FAILURE PROTOCOLS:
If behind at Day 30: Pivot to cost-cutting tech
If behind at Day 60: Focus on defensive moves
If behind at Day 75: Prepare acquisition readiness
Level 3 shows a thorough strategy - multiple angles considered, every stakeholder mapped, multiple scenarios planned. The three pillars are fully integrated: Empathy (understanding all stakeholders), Critical Thinking (failure protocols), and Human In The Loop (adaptive responses). Note: as usual 'Trajectory' planning requires all three pillars to be used iteratively and adapt as you go along for best results.
The Customer Crisis Pattern: Validation Techniques
How to handle high-stakes customer crisis situation, a bad and good example then Level 1 to Level 3 examples of validation techniques:
Bad Example: Respond to customer complaint
Good Example: [OBJECTIVE] Turn viral complaint into brand win [NEEDS] 50K retweets, valid issue, CEO watching [HOW] Crisis manager mindset - empathetic but strategic [TRAJECTORY] Acknowledge ? Explain ? Fix ? Follow-up
This approach helps build sophistication while maintaining clarity.
Validation Checkpoints: From Basic to Bulletproof Build checks into your Trajectory with increasing layers of sophistication.
Bad Validation: "Make sure this is good"
Good Validation: "Confirm financial calculations, verify customer quotes are accurate, check political feasibility with stakeholder perspectives, rate confidence 1-10 with reasoning" This good version provides specific checks across multiple dimensions. Let's look at how we can layer in sophistication and adapt.
[VALIDATION PROTOCOL]
STAGE 1 - Logic Validation:
"Before proceeding, confirm:
- Do conclusions follow from data?
- Are assumptions explicitly stated?
- Would a hostile audience find gaps?
- Where might black swans hide?"
STAGE 2 - Board Pre-Flight:
"For each recommendation, verify:
- CFO lens: Is ROI calculation bulletproof?
- CEO lens: Does this tell a compelling story?
- Board lens: How does this compare to best-in-class?
- Team lens: Is this actually executable?"
STAGE 3 - Stress Testing:
"Challenge mode:
- What if budget cut by 50%?
- What if key team member leaves?
- What if competitor moves first?
- What if market conditions worsen?"
STAGE 4 - Reality Calibration:
"Confidence scoring with evidence:
- High confidence (80%+): Based on what data?
- Medium confidence (50-79%): What would increase this?
- Low confidence (<50%): Should we even propose this?
- Assumptions: List all, rate impact if wrong"
STAGE 5 - Meta-Validation:
"Step back and ask:
- Are we solving the real problem?
- Does this plan anticipate competitor moves and actions?
- Would I bet my job on this?
- What question should we be asking instead?"
Common Pitfalls: Intermediate Level
Pitfall 1: Over-Engineering
Level 1 Mistake: Adding complexity without purpose Level 2 Mistake: Using every framework you know Level 3 Mistake: Creating such sophisticated prompts they become unwieldy
Fix: Every element must earn its place by improving outcomes
Pitfall 2: Ignoring Audience Psychology
Level 1 Mistake: Focusing on accuracy over reception Level 2 Mistake: Understanding audience but not their pressures Level 3 Mistake: Mapping politics but forgetting humanity
Fix: Remember that humans make decisions emotionally, then tend to justify rationally
Pitfall 3: Single-Pass Thinking
Level 1 Mistake: Expecting one perfect output Level 2 Mistake: Iterating without purpose Level 3 Mistake: Over-iterating into analysis paralysis
Fix: Design iteration cycles - each with specific improvement goals
Pitfall 4: Forgetting the Three Pillars
Empathy Gaps: Focusing on logic while ignoring human impact Critical Thinking Gaps: Accepting AI's first response without challenge HITL Gaps: Treating AI as autonomous rather than being a collaborative partner
Fix: Consciously apply all three pillars throughout your interaction
Integration: How Everything Works Together
The magic happens when ONHT and the three pillars work as one system:
O - Objective sets the strategic context (what success looks like) N - Needs provides the information architecture (what matters and why) H - How defines the thinking approach (mental models and expertise) T - Trajectory guides the journey (step-by-step with validation)
Throughout this structure:
Empathy ensures human understanding at every level
Critical Thinking challenges assumptions and validates logic
Human In The Loop (HITL) keeps you actively engaged in guiding and refining
This isn't just prompt engineering. It's intelligence augmentation that compensates for what AI naturally lacks.
Practice Challenges: Build Your Skills
Challenge 1: The Board Crisis
Scenario: Revenue down 20%. Board meeting tomorrow. CEO needs answers.
Level 1 Task: Create basic slides about revenue decline Level 2 Task: Design presentation addressing board concerns Level 3 Task: Architect complete board strategy including pre-meeting influence, presentation dynamics, and follow-up plan
Apply ONHT with all three pillars. Notice how your approach changes at each level.
Challenge 2: The Technical Migration
Scenario: Legacy system to cloud. 50 dependencies. Team nervous.
Level 1 Task: List steps for cloud migration Level 2 Task: Create migration plan with timelines and risks Level 3 Task: Design migration strategy that satisfies tech team capabilities, finance constraints, business continuity, and political dynamics
Focus on how Empathy, Critical Thinking, and HITL guide your trajectory.
Challenge 3: The Customer Rescue
Scenario: Major client threatening to leave. Worth 15% of revenue.
Level 1 Task: Draft retention offer Level 2 Task: Create multi-touchpoint retention strategy Level 3 Task: Architect a relationship transformation plan that turns a critic into advocate while creating replicable playbook
Each level demands deeper application of the framework and pillars.
What Makes You Valuable
Remember why you're learning this. Tools will change. Companies will change. But professionals who can:
Structure complex problems clearly
Guide AI to exceptional outputs
Navigate political and practical realities
Bridge the gap between AI logic and human needs
Iterate strategically to complete solutions
They'll always be valuable. Anywhere.
Start now to learn and excel in the ONHT framework and hone your skills in Empathy, Critical Thinking, and Human In The Loop
Your Growth Edge
You're no longer just getting outputs. You're designing interactions.
Each prompt teaches you about:
Information architecture
Cognitive patterns
Strategic communication
Political navigation
Systems thinking
These meta-skills compound. Fast.
Key Intermediate Principles
Layer objectives to reveal complete context - strategic, political, and human
Structure information for how executives actually make decisions
Stack thinking approaches to create sophisticated analysis frameworks
Design trajectories that adapt based on discoveries
Build validation that catches both logical and political failures
Always apply the three pillars: Empathy, Critical Thinking, and HITL
Recognise you're compensating for AI's missing cognitive functions
Your Next Evolution
Master Level 3 consistently before moving to advanced. When you can:
Map all stakeholders and their agendas instinctively
Structure information for immediate executive decision-making
Design adaptive trajectories that handle any discovery
Build validation that would satisfy your harshest critic
Navigate political dynamics while solving real problems
Seamlessly integrate the three pillars into every interaction
Then you're ready for advanced techniques.
Critical Insight
The best intermediate users develop intuition for which level of sophistication each situation needs. Not every problem needs Level 3 - but knowing when it does makes you indispensable.
Remember: You're not just using AI. You're creating a hybrid intelligence system that combines the best of human and artificial thinking.
The gap between intermediate and advanced isn't about more techniques. It's about seeing the system behind the techniques.
You're almost there.
Note: This framework is living documentation. It improves when you use it and share what works. The combination of ONHT structure and the three pillars creates something greater than the sum of its parts: a thinking partnership that solves problems neither human nor AI could handle alone.
In this Series: Introduction to the ONHT Framework The ONHT Framework: Beginners Guide The ONHT Framework for Intermediate users Coming Next: The ONHT Framework for Advanced users (WIP) Some ONHT Tips and Supporting Prompts (WIP) Key Advanced Features of the ONHT Framework (WIP)
This Intermediate Guide for the ONHT (Objective, Needs, How, Trajectory) Framework transforms you from someone who uses GenAI into someone who thinks with GenAI by adding the missing cognitive functions that current GenAI lacks. The framework works through three critical pillars – Empathy (understanding all stakeholders), Critical Thinking (challenging assumptions), and Human in the Loop (active partnership). Master these patterns and you'll be solving complex problems others can't even approach, becoming indispensable by designing interactions that produce exceptional results rather than just functional outputs.
Stop getting generic AI responses. Learn the four-letter framework that transforms vague requests into precise results. The ONHT framework: Objective (what problem you're solving), Needs (key information that matters), How (the thinking approach), and Trajectory (clear steps to the answer), teaches you to think WITH AI, not through it, turning "analyse customer feedback" into board-ready insights. Real examples show how adding context and structure gets you from Level 1 basics to Level 3 mastery, where AI delivers exactly what you need.
The difference? Knowing how to ask.
GenAI tools are transforming work, but most people get poor results because they don't understand how to communicate with AI built on structured data. This guide is a series of articles that teaches the ONHT framework—a systematic approach to prompting that transforms vague requests into exceptional outputs by focusing on Objectives (what problem), Needs (what information), How (thinking approach), and Trajectory (path to solution). Master this framework and develop an expert mindset grounded in human-in-the-loop thinking, critical analysis, and empathy, and you'll excel with any AI tool, at any company, in any role.
I question the reliability and accuracy of Generative AI (GenAI) in enterprise scenarios, particularly when faced with adversarial questions, highlighting that current Large Language Models (LLMs) may be data-rich but lack in reasoning and causality. I would call for a more balanced approach to AI adoption in cases of assisting users, requiring supervision, and the need for better LLM models that can be trusted, learn, and reason.
