Ertuğrul Çavuşoğlu — Product & AI Leader

Mindlid — Case Study

Role: Co-founder & CEO (Jan 2025– ) · Started Early 2023 (Ideation)

One‑liner: Non‑clinical emotional‑wellness companion featured by Google 4× across Gemini product coverage and developer showcases.

The bet: The wedge is emotional support. The prize is owning the personal memory layer for what HBR ranks the #1 consumer use of GenAI.

Origin: When ChatGPT made AI conversational, the obvious question was what it could do. I chased a harder one: could it go deep, carrying real emotional and cognitive weight, not just surface tasks? An early prototype, tested with a practicing psychologist, gave the first signal it could. That became Mindlid, and the traction below is it playing out.

Context & Team

• HBR's 2025 usage study later ranked companionship and therapy the #1 consumer use of GenAI, the space Mindlid was already building in.
• Assembled founding team across AI, product, design, and mobile (senior contributors).
• Two strategic pivots to a differentiated non‑clinical scope with clear guardrails.

Problem → Bet

• People want an accessible, private way to feel seen, relieved, and hopeful—without clinical claims.
• Bet: structured, therapy‑inspired programs + tone customization + couples mode can reach mass adoption if reliability and cost stay in check.

Constraints

• Investor‑safe: privacy, safety, non‑clinical language.
• Cost/latency targets per route; eval gates before shipping.
• Small team, rapid iteration cadence.

What I led

• Product vision & operating system; roadmap, success metrics, and GTM loops.
• Structured program design, guided narration, context continuity, and emotional‑wellness user journeys.
• Owned Mindlid’s product and AI operating layer across roadmap, evals‑as‑QA, reliability, safety/guardrails, analytics, bilingual AI execution, and OKR alignment.
• Led an agentic prompt/workflow workbench that cut prompt‑iteration‑to‑deploy from months to under 1 hour.

Key Decisions

• Non‑clinical scope with crisis resources and refusal policies.
• “Outcome before output:” A→O tests weekly.
• Context discipline + caching to keep CPT low; validator/formatters for safety.

Results & Signals

The bet, in the data: Mindlid is a repeat-usage product, so retention and session depth are the thesis proving out.

For context: peer-reviewed research on mental-health apps reports a median 4% daily open rate and 3.3% 30-day retention (JMIR, 2019); consumer health & fitness apps average ~2.5 minutes in-app per day (Apptopia).

• Featured by Google 4× across Gemini product coverage and developer showcases, including Gemini Embedding, Gemini Embedding 2, Gemini 3.1 Flash TTS, and a Google AI Developers showcase. Founder quoted in the Embedding 2 and Flash TTS launch posts alongside executives from Paramount Skydance, Everlaw, Sierra, and HeyGen.
• Mindlid’s core asset is its personal memory layer: cross-modal memory across text, audio, and visual that compounds with every conversation, making the product harder to leave the longer you use it. Cited in Google’s Gemini Embedding 2 launch announcement (+20% lift in top‑1 recall).
• Ran a trial‑gated, two-month alpha that reached 46 countries with no paid acquisition; engaged alpha users averaged ~1h 24m active use per day in a 3-day trial (vs the ~2.5 minute category‑wide average above), with ~50% of DAU returning on active days and WAU/MAU in the ~24–40% range. Rebuilt and relaunched as beta in 2026; shipped v1 in July 2026.
• Demonstrated early willingness‑to‑pay through voluntary pre‑launch subscription revenue, without a formal monetization push.
• Led structured user research with interviews and surveys; ~90% indicated willingness to pay based on the concept and prototype.
• Built early GTM momentum through revenue‑share creator affiliates, with ~2M combined reach across social, SMS, YouTube, website placements, and Telegram/WhatsApp distribution.
• Recognized alongside major brands: App Growth Summit Istanbul 2026 listed Mindlid among “powerhouse brands joining” the event, with Coca-Cola, Trendyol, Unilever, iGA Istanbul Airport, Boyner, Lyft, and others.

Refs: HBR: How people are really using GenAI (2025) · Google Developers Blog: Gemini Embedding · Google Blog: Gemini Embedding 2 · Google Blog: Gemini 3.1 Flash TTS · Google AI Developers showcase · Google DeepMind: Gemini Embedding models page (Mindlid showcase quote) · Google DeepMind: Gemini speech generation page (Mindlid showcase quote)

Back to Work

Emma Up — Case Study

Role: Head of Product & Technology (Founding Team) · Emma — The Sleep Company (2023–2024)

One‑liner: Apple‑featured sleep & habit coach that reached its first 100k users (CEO-stated); covered by T3/Yahoo; launch features via Reuters Plus & Yahoo Finance.

Context & Team

• Led Backend, AI, QA Automation, Flutter, and Design inside a 1,000‑employee org, the world's largest D2C sleep brand ($1B+ revenue in 2023).
• Coaching program developed with sleep experts from Oxford University; sound‑based tracking approach shown more precise than physical trackers in a Stanford study (per Emma's official launch release).
• Rebuilt the app to PMF; leveraged high‑traffic web and paid channels.

Problem → Bet

• Fragmented sleep advice didn’t translate into sustained behavior change.
• Bet: combine passive signals + habit coaching + crisp onboarding to drive retention and revenue.

What I led

• End‑to‑end redesign and growth systems (onboarding, paywall, retention loops).
• LLM features analyzing 19 sleep & habit parameters for personalized challenges.
• Integration with Asleep (Korea) for hands‑free, sound‑based tracking.
• Wearables data ingestion via native iOS and Android health integrations, with Spike API covering long‑tail devices, including rings.
• Data plumbing, instrumentation, and experiment cadence (weekly).

Key Decisions

• Checkout integration for direct EBIT impact.
• “First value in minutes” onboarding; reduce cognitive load and steps.
• Price/plan experiments aligned to habit momentum.

Results (selected)

• ~€50M A/B-attributed EBIT impact via checkout integration; the A/B tested checkout with the app versus without, so the measured delta is the app's effect on the parent P&L (growth-team readout).
• 7.8× improvement in 6‑week retention (qualified users).
• 20× improvement in Month‑3 subscription retention vs prior year (named in the CEO-signed reference letter, Dec 2024).
• €500k+ internal funding secured to scale the AI Sleep Coach.

Demos: User Testimonial (12k views) · Launch Video

Attested in writing: the official company reference letter (Co-CEO-signed, Dec 2024) credits co-founding and developing Emma Up v2 and leading the initiatives behind the EBIT increase; a signed confirmation (June 2026) by Felix Focken, former Global Head of Emma Up, credits the product roadmap and business case behind the €500k+ internal funding; his public LinkedIn recommendation (Aug 2026) further attributes an eight-figure EBIT impact to the checkout integration, per the growth team's A/B readout. Both letters available on request.

Back to Work

UserGuiding — Case Study

Role: Product Manager (2020–2021)

One‑liner: Drove adoption and growth; Top‑5 Product Hunt ×2.

Context

• Mid‑market SaaS with self‑serve onboarding.
• Needed higher feature discovery and stickiness without heavy services.

Problem → Bet

• Users weren’t finding help at the moment of need.
• Bet: Resource Centers + guided prompts + better analytics would lift activation and usage.

What I did

• Discovered & shipped Resource Centers (contextual help hub).
• Built analytics with BigQuery and PostgreSQL; instrumented core features.
• Launched Covid Meter and refined guided prompts for timely value.

Results (selected)

• 60% adoption of Resource Centers in 2 months (beta).
• Top‑5 Product Hunt ×2.
• +40% engagement on core features.
• ARR +500% (~10% MoM); supported a $1M seed round.
• UserGuiding entered G2's Leader quadrant in Spring 2021, during this tenure, and has held it for 20 consecutive quarters since; in 2026 it is the only Digital Adoption Platform on G2's Best Software list (top 0.6% of 175,000+ products).

Back to Work

EMnify — Case Study

Role: Growth Product Manager (2022–2023)

One‑liner: PLG improvements across onboarding and evaluation flows.

Problem → Bet

• Activation friction and time‑to‑value blocked growth.
• Bet: checklists + interactive help + prepaid/eval flow redesign would speed value and lift conversion quality.

What I did

• Onboarding Checklists v1 + Interactive Help.
• Prepaid balance & evaluation card flow redesign (shortened steps, clearer value).
• Integrated PQL initiatives with Salesforce.

Results (selected)

• Activated users ↑43% in 3 months; new daily record for evaluation orders.
• Time‑to‑value ↓90% (1h25 → 9m).
• Incentives campaign drove 54% of evaluation orders; signup→activation ↑20%.

Back to Work