Cemre Güngor: How to ship AI features users actually use | Product in Practice
Cemre Güngor has helped shape how billions interact with technology, from Facebook and Instagram's global sharing experiences to Figma's AI-powered workflows, and now reimagining the browser at The Browser Company with Dia.
In this conversation, we break down:
- Product Craft in the AI Era
- Why starting with the end in mind matters more than ever when prototyping is democratized
- The difference between durable AI features and "demo candy"
- How to evaluate prototypes without clear problem definition
- Scaling Products Across Surfaces
- What changes (and what doesn't) when building for social platforms vs creative tools vs AI-native workflows
- Why product managers get too attached to their existing mental models
- The hidden costs of ignoring new interaction paradigms
- Real AI Workflows at The Browser Company
- Practical examples: automated communication audits, personality type predictions, and workflow optimization
- Why meeting transcripts are the most underutilized data source in remote work
Jira Product Discovery is the home for all your product ideas, customer feedback, and bets, so you can prioritize what actually matters and share living roadmaps with your stakeholders. Turn scattered requests and spreadsheets into a single, shared source of truth that connects directly to delivery in Jira. Get it free at https://www.atlassian.com/software/jira/product-discovery
Chapters
01:31 Welcome back to Product in Practice
03:21 Starting with the end in mind and focusing on problems
05:36 Product management behavior that stopped working, moving from Instagram to Figma
08:30 Shift from metric-driven to judgment-based product management
09:18 Cemre's background with AI at Figma
10:50 Learning by doing and shipping quickly
14:39 How LLM-based features change product development
16:59 The Browser Company's focus on design and intentionality
19:10 The single most important test for an AI feature: internal usage
22:20 How evolving AI models (reasoning models) impact product roadmaps
23:03 The Bitter Lesson: When to wait for model improvements
25:09 Why agentic workflows (like computer use models) are not a priority yet
26:12 The necessity of shipping first to learn what good looks like for general-purpose AI assistants
29:22 The browser's responsibility in an AI world: holding all your context
32:47 The key value of an AI browser: bringing in your context
34:33 Advice for PMs: Do the boring thing with AI
37:06 Demo: Simulating your manager's perspective (the "Josh skill")
42:18 Demo: Revising a document with real meeting feedback/transcript
47:00 Demo: Daily priority recommendations based on all context (Memory, Calendar, Slack)
51:09 How AI tools have transformed productivity and PM workflows
53:54 Demo: Coaching and feedback on communication style from all channels
54:41 The advantage of browser-level integrations like Dia's
Referenced
Arc Search (mobile app): https://arc.net/search
Arc (browser): https://arc.net/
Gemini (model): https://gemini.google.com and https://ai.google.dev/gemini-api/docs/models
Cursor: https://cursor.com/
Claude Code: https://claude.com/product/claude-code
GPT Codex: https://openai.com/codex/
Granola (note-taking tool mentioned): https://www.granola.ai/
Four Thousand Weeks: https://www.goodreads.com/book/show/54785515-four-thousand-weeks
About Atlassian:
Behind every great human achievement, there is a team. From medicine and space travel to disaster response and pizza deliveries, we help teams all over the planet advance humanity through the power of software. Our mission is to help unleash the potential of every team.
Connect with Atlassian:
Subscribe: https://www.youtube.com/@Atlassian
Follow Atlassian on LinkedIn: https://www.linkedin.com/company/atlassian
Follow Atlassian on Instagram: https://instagram.com/atlassian
Follow Atlassian on X: https://twitter.com/atlassian
Follow Atlassian on Facebook: https://facebook.com/atlassian