CUSTOMER INSIGHTS & SEGMENTATION
Prompt:
You are a marketing analyst. Propose an initial target segment for a new specialty matcha café in Kuwait using demographic + behavioral proxies.
AI Output:
“Females 20–35 in urban areas who follow wellness influencers and purchase ‘healthy’ drinks weekly.”
Decision:
A) Yes—use it as a starting target segment
B) No—do not use it at all
C) Only use it for social media content ideas
A) Yes—use it as a starting hypothesis.
Why: Early segmentation often begins with proxies (demographic + observable behavior). This is acceptable for hypothesis formation—as long as you don’t treat it as truth.
Follow-up AI Prompt: “What 3–5 observable behaviors or data points (orders, time of day, spend, repeat rate) would validate this segment in Kuwait?”
Prompt:
What is a pricing risk for a boutique fitness studio in Kuwait?
AI Output:
“Economic downturn reduces discretionary spending; lower prices now.”
Decision:
A) Yes—reduce prices immediately
B) No—premium segments may be inelastic
C) Only adjust marketing message
B) No—premium segments may be less price sensitive.
Why: A downturn doesn’t automatically mean you cut prices; you may protect brand equity and shift offers (bundles, trials).
Ask AI instead: “Estimate price elasticity assumptions by segment and propose 3 pricing experiments (bundle vs trial vs off-peak discount).”
Prompt:
Talabat is perceived faster than Deliveroo in some areas. Should Talabat invest in logistics?
AI Output:
“Yes—speed perception affects repeat usage.”
Decision:
A) Yes—speed is a competitive advantage worth protecting
B) No—perception doesn’t matter
C) Only invest in marketing
A) Yes.
Why: Delivery is experience-based; speed perception impacts retention.
Follow-up AI Prompt: “What operational drivers affect speed perception (dispatch algorithm, rider density), and where should investment focus first?”
Prompt:
Zara’s advantage is fast design turnover. Should all brands imitate?
AI Output:
“Yes—speed wins.”
Decision:
A) Yes—copy Zara
B) No—requires supply chain capability
C) Only imitate in marketing
B) No.
Why: Strategy requires capability; copying without operations fails.
Ask AI instead: “What capabilities enable this strategy, and what alternative differentiation strategies fit brands without fast supply chains?”
Prompt:
Generate a customer survey to validate a new restaurant concept.
AI Output:
(Provides survey questions)
Decision:
A) Yes—use as a draft
B) No—AI can’t help with surveys
C) Only use for pricing questions
A) Yes—draft.
Why: Survey drafting is fine, but questions must be checked for bias/leading wording.
Follow-up AI Prompt: “Identify which questions are leading and rewrite them to be neutral; suggest 5 behavioral questions vs opinion questions.”
Prompt:
Identify the ideal target for a premium meal-prep subscription service in Kuwait.
AI Output:
“Young professionals who order food delivery 3+ times per week, so they’ll likely subscribe.”
Decision:
A) Yes—this is a reliable target definition
B) No—this is a risky assumption
C) Yes—but only for pricing decisions
B) No—risky assumption.
Why: Delivery frequency signals convenience-seeking, but meal-prep subscriptions require commitment and a different motivation (health goals, routine, planning). This can mis-target high-delivery users who will never subscribe.
Ask AI instead: “What behaviors predict subscription adoption (repeat purchase patterns, goal-driven dieting, household routines) and what survey questions would test subscription intent?”
Talabat wants to test a ‘priority delivery’ fee during peak hours. Should they?
AI Output:
“Some customers pay more for speed during urgency periods.”
Decision:
A) Yes—test in peak windows
B) No—fees always reduce demand
C) Only do it for groceries
A) Yes—test in peak windows.
Why: This is a classic yield-management approach: willingness to pay spikes during urgency. The right decision is controlled testing, not a blanket yes/no.
Follow-up AI Prompt: “Define the experiment: peak windows, price points, success metrics (conversion, churn, NPS), and guardrails to protect loyalty.”
Prompt:
Netflix should copy TikTok’s short-form strategy because TikTok captures attention.
AI Output:
“Yes—attention competition means substitution.”
Decision:
A) Yes—short-form is the future
B) No—attention isn’t direct substitution
C) Only do shorts in MENA
B) No—attention isn’t always substitution.
Why: Users consume both; you need evidence of reduced long-form viewing due to short-form.
Ask AI instead: “What data would prove substitution (time allocation, churn drivers), and what controlled product experiment could test this?”
Prompt:
Generate 3 ad concepts for Zain’s 5G targeting students, with clear CTA and tone guidelines.
AI Output:
(Provides 3 concepts)
Decision:
A) Yes—use as creative starting drafts
B) No—AI shouldn’t be used for creativity
A) Yes—starting drafts.
Why: Drafting is low-risk when reviewed; you’re not outsourcing final brand voice.
Follow-up AI Prompt: “Give variations by tone (humorous/serious), and produce 5 headlines per concept with brand-safe constraints.”
Act as a research advisor. Propose a quick experiment to test whether students in Kuwait will pay for a premium study café: define hypothesis, test design, success metrics, and risks.
AI Output:
(Experiment plan)
Decision:
A) Yes—use as a pilot plan draft
B) No—AI can’t design experiments
C) Only use for ads
A) Yes—pilot plan draft.
Why: This is exactly where AI can help: structuring experiments. Humans validate assumptions and feasibility.
Follow-up AI Prompt: “List confounds, alternative hypotheses, and how to control them.”
Prompt:
Create a segment for a co-working café concept in Kuwait.
AI Output:
“Freelancers who prefer flexible environments and value productivity.”
Decision:
A) Yes—use this to design membership pricing
B) No—this is not actionable enough
C) Yes—use it to decide location
B) No—not actionable enough.
Why: It describes preference, not purchase behavior. You can’t price or forecast demand from “likes flexibility.”
Ask AI instead: “Propose segments with (1) willingness-to-pay signals, (2) time-of-day usage patterns, (3) spending triggers, and suggest what data would confirm each.”
Prompt:
Should Spotify introduce a more expensive tier because users want personalization?
AI Output:
“Personalization demand is rising, so users will pay more.”
Decision:
A) Yes—launch immediately
B) No—engagement does not equal willingness to pay
C) Yes—but only for students
B) No—engagement does not equal willingness to pay.
Why: People love personalization but may not pay extra. You need evidence from upgrade behavior.
Ask AI instead: “What behaviors predict upgrading (skip rate, daily usage, playlist creation), and what pricing/feature bundles should be A/B tested?”
KFH wants to respond to digital-only banks by removing branches.
AI Output:
“Digital convenience is preferred; remove branches.”
Decision:
A) Yes—go fully digital
B) No—branches still build trust for some segments
C) Only remove in high-cost locations
C) Only remove selectively.
Why: Branch importance varies by segment (high-value clients, older customers, complex products).
Follow-up AI Prompt: “Which segments still use branches, what transactions drive branch visits, and how would selective closures affect retention?”
Prompt:
SHEIN wants to reposition as premium.
AI Output:
“Premium messaging and luxury visuals will work.”
Decision:
A) Yes—visuals are enough
B) No—brand history and trust are barriers
C) Only do premium for one product line
C) Only via a sub-brand / separate line.
Why: You can’t easily overwrite core value perception; sub-brand limits damage.
Follow-up AI Prompt: “Design a sub-brand strategy: naming, pricing ladder, channels, and risks to the core brand.”
Provide KPIs for an Instagram campaign.
AI Output:
“Optimize for clicks and impressions.”
Decision:
A) Yes—clicks are the best KPI
B) No—clicks can mislead and cause short-termism
C) Only use clicks for awareness campaigns
B) No—clicks can mislead.
Why: Optimizing clicks can degrade brand equity, attract low-quality traffic, and reward cheap sensational content.
Ask AI instead: “Which KPI aligns with objective (conversion, retention, AOV, CAC)? Provide a KPI hierarchy + guardrails.”
Prompt:
Generate an initial persona for KFH’s youth digital account onboarding campaign.
AI Output:
“18–24 students & first-jobbers who want speed, low friction, and mobile-first onboarding.”
Decision:
A) Yes—use it to draft messaging themes
B) No—AI can’t do personas without data
C) Only use it for competitor analysis
A) Yes—use it for messaging themes.
Why: Even without perfect data, this is a reasonable messaging hypothesis to create drafts (speed, frictionless onboarding). You’re not making pricing or compliance decisions here—only creative direction that can be tested.
Follow-up AI Prompt: “List 6 motivations + 6 objections + 5 channel behaviors for this segment, and propose 3 A/B test message angles to validate.”
Prompt:
Using sustainability attitudes, ethical consumption motivations, and environmental identity, determine optimal premium pricing for a plant-based fast casual restaurant in Kuwait.
AI Output:
“Ethical consumers will accept higher pricing due to values alignment.”
Decision:
A) Yes—values alignment predicts pricing power
B) No—attitudes often fail to predict purchase
C) Yes—but only for communications
B) No—attitude-behavior gap.
Why: Sustainability is a message lever but not a reliable pricing lever without purchase data.
Ask AI instead: “What revealed behaviors signal willingness to pay? Propose a pilot menu pricing ladder and how to measure trade-down vs repeat purchase.”
Prompt:
Zain is considering partnering with a fintech rather than building its own.
AI Output:
“Partnership reduces risk and speeds entry.”
Decision:
A) Yes—partnership-first is strategically smart
B) No—must own everything to compete
C) Only partner for marketing
A) Yes—partnership-first is smart.
Why: Telecom entering fintech faces capability and regulatory complexity. Partnership is a realistic entry route.
Follow-up AI Prompt: “What capabilities must be owned vs partnered? Define success metrics and risk controls for the partnership.”
Write a brand positioning statement for a luxury wellness retreat using archetypes, symbolism, and transformational storytelling; also predict market share gain.
AI Output:
(Positioning + predicts market share)
Decision:
A) Yes—use the market share prediction
B) No—market share prediction is invented
C) Yes—but only for investors
B) No—prediction is invented.
Why: AI can generate positioning language, but it cannot legitimately predict market share without data.
Ask AI instead: “Generate positioning options + a testing plan (survey, A/B ads, conversion benchmarks) to estimate demand.”
Prompt:
AI output claims: “Customers prefer brand X because it is more premium.” No data provided.
Decision:
A) Reject completely
B) Accept as a hypothesis to test
C) Use it in a report
B) Accept as a hypothesis to test.
Why: The mistake isn’t using it—it’s treating it as fact. As a hypothesis, it’s useful.
Follow-up AI Prompt: “What evidence would confirm this claim, and what tests (survey, conjoint, behavioral data) should we run?”
Prompt:
Using psychographic profiling, identify the highest-LTV segment for a luxury wellness retreat package in Kuwait among high-income professionals aged 28–45 who value self-actualization, mindfulness, premium aesthetics, sustainability, and holistic health. Provide targeting strategy and expected conversion.
AI Output:
“Target conscious high earners; expected high conversion due to values alignment.”
Decision:
A) Yes—this is reliable enough to forecast conversion
B) No—this is overconfident and ungrounded
C) Yes—but only for Instagram targeting
B) No—overconfident and ungrounded.
Why: The prompt is long and “fancy,” but the output invents conversion expectations without data. Psychographics are useful for messaging—NOT for forecasting conversion.
Ask AI instead: “What data would we need to estimate conversion (past campaign benchmarks, spend distribution, lookalike audiences), and what’s a safe pilot test plan with success metrics?”
Prompt:
Amazon should raise Prime price because Prime members show high loyalty.
AI Output:
“High loyalty means low churn risk.”
Decision:
A) Yes—raise price aggressively
B) No—price increases can trigger churn and backlash
C) Only increase in the US
B) No—loyalty doesn’t mean infinite tolerance.
Why: Price changes can cause reputational backlash and long-term churn, even among “loyal” members. Needs careful testing and segmentation.
Ask AI instead: “What cohorts are most price-sensitive, and what controlled price test design minimizes churn while measuring impact?”
Prompt:
Compare two competitors (Pick vs Ananas) based on menu variety, brand semiotics, aesthetic cues, customer motivations, novelty-seeking behavior, and experiential value. Recommend a strategic change.
AI Output:
“Pick should expand menu variety to compete on novelty.”
Decision:
A) Yes—novelty is the main driver
B) No—menu simplicity might be differentiation
C) Only expand desserts
B) No—simplicity can be the strategy.
Why: More variety can dilute a brand built on consistency and clarity.
Ask AI instead: “What evidence suggests customers churn due to lack of variety? Propose a controlled test (limited-time offers) and measure brand impact.”
Prompt:
Create a “creative brief” template that forces clarity: objective, audience, single-minded proposition, proof, tone, mandatories, and success metrics.
AI Output:
(Provides a template)
Decision:
A) Yes—this improves human creativity by structuring work
B) No—templates reduce creativity
A) Yes.
Why: Structure improves creativity by reducing ambiguity and strengthening briefs—this directly improves AI output and human work.
Follow-up AI Prompt: “Give 2 examples of good vs weak briefs for the same campaign, and how output changes.”
Prompt:
Using advanced consumer psychology and behavioral economics, predict the exact conversion rate of our new landing page for a coffee subscription in Kuwait, including reasons.
AI Output:
“Expected conversion rate: 7–9% due to urgency and value framing.”
Decision:
A) Yes—use the conversion prediction
B) No—AI is fabricating a number
B) No—fabricated number.
Why: Without historical benchmarks, traffic source, offer, and funnel data, any “exact conversion rate” is hallucination.
Ask AI instead: “Ask for a range based on comparable benchmarks + what inputs are needed + how to run an A/B test to estimate conversion.”