AI Transformation at Scale: Practical Adoption Across 150+ Engineers with Peter Gostev // Head of AI @ Moonpig
Navigating AI hype versus implementation reality in large engineering organizations
Peter Gostev · Head of AI
How do you drive meaningful AI transformation across 150 software engineers without mandates or force? Peter Gostev, Head of AI at Moonpig, reveals the technical strategies and organizational approaches behind scaling AI adoption from 130 to 400+ users while navigating the gap between industry hype and implementation reality.
From managing complex integration challenges where building AI-powered features requires 80-90% traditional software engineering (integrations, scaffolding, deployment, UI) to implementing three-pillar strategies (tool adoption, automation workflows, experimental features), Peter shares hard-earned insights on building AI capabilities through process re-engineering rather than simple automation.
Technical insights for CTOs and engineering leaders:
- 🏗️ Portfolio approach: balancing quick wins with experimental high-impact projects
- ⚡ Prototype-first methodology for validating AI solutions before full development
- 🤖 Reality gap between agentic AI hype and production deployment complexity
- 👥 Organic adoption strategies that scale without top-down mandates
- 🔧 Custom GPT frameworks for non-technical subject matter experts
- 📊 Why most AI work is integration, scaffolding, and deployment—not just AI
- 🔄 Process re-engineering with AI: changing workflows rather than automating existing inefficiencies
Show notes.
The AI transformation challenge isn't just technical—it's organizational. Peter Gostev, Head of AI at Moonpig, discusses building practical AI capabilities in a 600-person company while navigating the gap between AI hype and implementation reality.
From managing adoption across 150 software engineers to partnering with non-technical teams, Peter shares insights on driving meaningful AI transformation that goes beyond superficial tool usage. This conversation explores the contrast between industry claims of widespread AI adoption and the nuanced reality of implementing AI at scale.
TIMESTAMPS: [00:00:51] Introduction: Peter Gostev, Head of AI at Moonpig [00:03:49] First Computer Memory: Pentium 2 and the Golden Age of Learning [00:05:28] Career Evolution: From Gaming to Data Science and Machine Learning [00:07:01] AI Hype Cycle: Underestimating Future Potential While Overhyping Present [00:09:41] Agent Reality Check: Why Full Agentic Systems Are Harder Than Expected [00:13:51] The Decade of Agents: Learning from Autonomous Driving Timelines [00:15:09] Developer Tool Adoption: The Skill Gap in AI-Assisted Programming [00:18:40] Organizational Change: Teaching 150 Engineers New Workflows [00:20:51] Claude Code Success: Finding Product-Market Fit in CLI Tools [00:24:33] Context Management: Why Proper AI Usage Requires Intuition [00:26:07] Three-Pillar AI Strategy: Tools, Automation, and Experimentation [00:28:01] Team Structure: AI Engineers vs Data Scientists [00:30:10] Partnership Models: Building vs Enabling Other Teams [00:32:51] Adoption Management: Direct Collaboration vs Ambassador Networks [00:35:45] Prototype Power: One Evening of LLM Work vs Weeks of Traditional Development [00:38:02] Subject Matter Expertise: Custom GPTs for Non-Technical Users [00:42:45] Writing with AI: Research vs Synthesis and the Importance of Human Thinking [00:47:08] Tool Strategy: ChatGPT Rollout Without Internal Marketing [00:49:58] Adoption Reality: 400 Users But Still Only 5-10% of Potential Usage [01:02:06] Training Challenges: From Individual to National Level Education [01:05:10] Future of Development: What Changes and What Remains Human [01:08:24] Developer Productivity: Higher Ambition, Not Less Work [01:16:50] University Concerns: Are Students Prepared for AI-First Development? [01:18:47] Career Advice: Do Hard Technical Things Early While You Can
QUOTES: • "I think there's never really been more exciting time in my career versus now and every time we can we can just do more and more and more, which is just such a cool position to be in." - Peter Gostev [00:06:39] • "I think people really underestimating where we are and or how much more there is to go… we are really just at the beginning of these investments starting to come through." - Peter Gostev [00:07:01] • "On any given project, most of the work is actually non-AI work… most of it is integrations, scaffolding, deployment, UI, all of the things like that. And then the AI part of it is normally maybe 10, 20% of the actual work." - Peter Gostev [00:28:01]