Skip to content
Aakash GuptaAakash Gupta

The Product Delight Framework for AI PMs (How AI Products Like ChatGPT Win)

Nesrine Changuel, former Product Leader at Spotify and Google Meet, reveals the complete framework for engineering delight into AI products. She breaks down the difference between surface and deep delight, the 4-step delight model, and why emotionally connected users are 2X more likely to stay, recommend, and buy. Full Writeup: https://www.news.aakashg.com/p/nasrin-shengel-podcast ---- Timestamps: 0:00 - Intro 2:02 - Surface Delight vs Deep Delight 4:58 - When Delight Goes Wrong: Apple Summary Disaster 11:04 - ChatGPT's Emotional Connection 15:55 - Ads 17:42 - Humanization: Compare to Human Service 24:13 - Google Meet's Delight Team & AI Translator 27:32 - 3 Types of Delight: Low, Surface, Deep 36:19 - Ads 38:18 - Gmail Smart Compose: Deep Delight Example 42:03 - B2B Delight: Jira & the "Superhero" Value 45:50 - The Delight Model: 4-Step Process 48:23 - Motivational Segmentation 55:05 - The Delight Grid & 50/40/10 Rule 1:02:39 - The Delight Checklist: 10 Questions 1:12:42 - Leaving Google to Solopreneur 1:18:45 - Outro ---- 🏆 Thanks to our sponsors: 1. Miro: The AI innovation workspace - https://miro.com/innovation-workspace/?irclickid=yIg1Kj2P2xycUXeyopwbUQf0UkpwPezrCXtgyg0&irgwc=1 2. Vanta: Leading AI compliance platform - http://vanta.com/aakash 3. Testkube: Leading test orchestration platform - http://testkube.io/ 4. Kameleoon: Leading AI experimentation platform - http://www.kameleoon.com/ ---- Key takeaways: 1. Deep delight beats surface delight - Confetti is surface delight. Deep delight is when functionality meets emotion (e.g., Gmail Smart Compose reduces stress while helping you write). 2. Emotionally connected users are 2X more valuable - Research from Capgemini, Deloitte, HBR, and McKinsey shows they stay longer, recommend more, and buy more vs "highly satisfied" users. 3. The Delight Model: 4 steps - (1) Identify user motivators (functional + emotional), (2) Turn motivators into product opportunities, (3) Create solutions, (4) Validate delight to avoid disasters. 4. Use motivational segmentation, not just demographics - Don't just ask "who are users?" Ask "why do they use your product?" Example: Spotify users want to search OR be inspired OR change their mood. 5. The Delight Grid maps features to motivators - If a feature only solves functional needs = low delight. Only emotional = surface delight. Both = deep delight. 6. Follow the 50/40/10 rule - Allocate 50% of roadmap to low delight (core functionality), 40% to deep delight (differentiation), 10% to surface delight (brand personality). 7. Humanization technique: Compare to human service - Google Meet compared features to "in-person meetings," not Zoom. Dyson compared robots to "hiring a human cleaner." 8. Corner cases kill delight in AI products - Apple's breakup message summary and WhatsApp's "ask John to resend it" (when John died) show why inclusiveness matters. Test 0.01% edge cases. 9. ChatGPT wins on emotional connection, not just accuracy - Users forgive inaccurate answers because of personalization, tone, and the feeling of "companionship" (especially for solo workers). 10. The Delight Checklist: 10 validation questions - Does it bring value to business? To user? Is it inclusive? Familiar enough? Continuous? Measurable? Use this before shipping. ---- 👨‍💻 Where to find Nasrin Shengel: LinkedIn: https://www.linkedin.com/in/nesrinechanguel/ Book: https://nesrine-changuel.com/product-delight-book/ 👨‍💻 Where to find Aakash: Twitter: x.com/aakashg0 LinkedIn: linkedin.com/in/aagupta/ Newsletter: news.aakashg.com #productdelight #aiproducts #productmanagement ---- 🧠 About Product Growth: The world's largest podcast focused solely on product + growth, with over 195K+ listeners. 🔔 Subscribe and turn on notifications to get more videos like this.

Aakash GuptahostNesrine Changuelguest
Nov 8, 20251h 19mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

How AI PMs engineer deep delight to build lasting loyalty

  1. The episode argues that winning AI products differentiate by engineering emotional connection alongside functional performance, not by shipping utility alone.
  2. It distinguishes surface delight (confetti, animations, Easter eggs) from deep delight (functionality designed to satisfy emotional needs) and warns that poorly designed “delight” can create harm and backlash.
  3. Multiple case studies—ChatGPT companionship, Google Meet’s emotion-preserving translation, Chrome’s Inactive Tabs, Gmail Smart Compose—show how deep delight increases trust, comfort, and loyalty.
  4. Changuel presents a four-step Delight Model (motivators → opportunities → solutions → validation) and emphasizes motivational segmentation as the starting point for product strategy.
  5. A prioritization heuristic (50/40/10) and a delight checklist operationalize delight work while keeping it aligned with business value, inclusiveness, and measurable outcomes.

IDEAS WORTH REMEMBERING

5 ideas

Deep delight is engineered, not decorated.

Surface delight adds emotional garnish (confetti, animations), while deep delight builds emotion into the functional experience itself—creating stronger attachment and differentiation.

Delight can backfire; inclusiveness and corner cases are part of the job.

Examples like Apple’s cold breakup summary, WhatsApp’s “ask them to resend it” grief moment, and Deliveroo’s fake missed-call campaign show how emotional misreads create disappointment and reputational damage.

Start with motivational segmentation: users share a product, not a “why.”

Changuel argues the best segmentation is motivational (the “why”), including both functional motivators (solve a problem, efficiency, ease) and emotional motivators (personal feelings and social identity/pride).

Humanization raises the bar more than competitor comparisons.

Teams at Dyson and Google Meet asked, “If this were done by a human, what would great service look like?”—leading to features like hand raise, reactions, and more “in-room” meeting behaviors.

Emotional connection outperforms satisfaction for growth outcomes.

She cites a consensus from major research (e.g., HBR/McKinsey/Deloitte/Capgemini) that emotionally connected users are ~2x more likely to retain, recommend, and buy more compared to merely ‘highly satisfied’ users.

WORDS WORTH SAVING

5 quotes

They’re not just better at their functional job. They’re doing something most AI product builders miss. They’re engineering delight.

— Aakash Gupta

It’s not about sprinkling confetti on top of utility. It’s about creating the product while addressing the emotional need at the same time.

— Nesrine Changuel

It’s better not to bring delight than to bring delight the wrong way.

— Nesrine Changuel

It’s not a feature. It’s a relationship that you need to build.

— Nesrine Changuel

The opposite of delight is disappointment.

— Nesrine Changuel

Functional vs emotional needs in productsSurface delight vs deep delight vs low delightCorner cases, inclusiveness, and “delight gone wrong”ChatGPT’s emotional connection and companionshipHumanization: comparing product experience to human serviceDelight Model (4-step process)Motivational segmentation (functional, personal emotional, social emotional)Delight Grid for roadmap mapping50/40/10 roadmap allocation ruleDelight checklist and validation metricsAnti-delight in freemium/upsell mechanicsB2B “B2H” delight and enterprise emotions (superhero, better leader)

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome