For more on building AI products and careers, along with early course announcement and special pricing, subscribe to the AI Career Boost mailing list at https://aicareerboost.com/interested

THE GUEST

Amri Abuseman is a powerhouse engineering leader known for building quality-first cultures across high-stakes industries—from healthcare and FinTech to telecom and HR tech. As Director of Engineering at Flatiron Health, she leads teams operating at the critical intersection of software delivery, system reliability, and regulatory rigor—where the margin for error is small and the impact of great engineering is enormous. With deep roots in quality engineering, Amri brings a powerful blend of technical depth and strategic leadership. Her toolkit spans everything from AI-driven test automation to enterprise-scale release management, helping organizations ship software that is not only fast, but reliable and resilient. But what truly sets her apart is her ability to align product and engineering through clear strategy, strong technical fluency, and genuine cross-functional empathy. Amri is the kind of leader who doesn’t just ship software—she builds the systems, cultures, and practices that make high-quality delivery sustainable at scale. She’s passionate about creating engineering environments where quality is embedded from the very beginning, not treated as an afterthought. And in this episode, we’re diving into what it really takes to build quality-first engineering organizations, how AI is reshaping test automation and software reliability, and why aligning product and engineering is the key to delivering software that truly matters.

THE SUMMARY

The fastest way to learn AI is to build, not study: Spending months watching tutorials or completing courses rarely leads to real capability. The only way to understand AI tools properly is to experiment, build rough prototypes, and learn through failure.

AI is removing the biggest barrier to building software: Modern AI coding tools are enabling people with little or no programming experience to create real products. This shift means the next wave of builders may come from non-traditional technical backgrounds.

AI tools only succeed if they actually fit developer workflows: Engineers quickly abandon tools that slow them down or misunderstand their code context. Real productivity gains only happen when AI tools integrate seamlessly with existing development habits.

Buying AI tools doesn’t guarantee productivity gains: Many organisations assume that simply adopting AI tools will instantly improve output. In reality, poor integration, unclear use cases, and workflow friction often lead teams to stop using them entirely.

Regulated industries are still pushing forward with AI innovation: Sectors like healthcare and finance face strict restrictions on using public AI systems. Instead of avoiding AI altogether, companies are building internal AI environments to stay competitive while maintaining compliance.

Product managers can’t afford to stay non-technical anymore: Leaders who avoid experimenting with AI tools risk losing visibility into what their teams can actually build. Getting hands-on with AI tools dramatically improves product intuition and decision-making.

The real AI skill isn’t prompting—it’s persistence: Success with AI often comes down to patience and experimentation. The builders who succeed are the ones willing to debug, refine prompts, and iterate repeatedly until something useful emerges.

THE SHOW

Weekly conversations with the AI’s top product leaders. Join Polly Allen as she discovers the paths to success in the world of AI.

THE LINKS

Have a question you want us to answer? Send it through to support@aicareerboost.com

Amri Abuseman

Linke

For more on building AI products and careers, along with early course announcement and special pricing, subscribe to the AI Career Boost mailing list at https://aicareerboost.com/interested

THE GUEST

Brian Diller is an AI product leader focused on turning complexity into clarity in higher education. At Watermark, he’s leading the design and launch of student success and course evaluation products, thoughtfully integrating AI into workflows that help institutions better support learners and make more informed decisions. With a rare blend of systems thinking and creative empathy, Brian brings a unique perspective to product leadership. Before stepping into the world of AI and product development, he spent years as a music professor—an experience that continues to shape how he approaches collaboration, problem-solving, and leadership today. Known for translating complex ideas into practical solutions, he focuses on building tools that are not only powerful, but genuinely useful for the people who rely on them. Brian is especially passionate about the responsible application of AI in real-world decision making—ensuring that emerging technologies support human judgment rather than replace it. And in this episode, we’re diving into how AI can meaningfully improve student success, the challenges of designing for higher education, and what it really takes to bring responsible AI into everyday institutional workflows.

THE SUMMARY

AI product leadership often starts with curiosity, not expertise: Getting involved in AI initiatives doesn’t require deep technical skills upfront. Asking to participate in projects, raising your hand early, and being willing to learn in public can quickly position you as the internal expert in emerging AI workflows.

AI works best as a thinking partner for product managers: Tools like Gemini and ChatGPT are incredibly effective for brainstorming product features, exploring competitive strategies, and refining ideas. Instead of replacing PM judgment, AI amplifies creative problem-solving and structured thinking during product discovery.

One of AI’s strongest use cases is synthesizing overwhelming data: Large lecture classes can generate hundreds of course evaluations, making manual analysis nearly impossible. AI can summarise patterns, detect recurring themes, and highlight actionable feedback, allowing educators to quickly understand what students are actually saying.

AI can transform fragmented student data into meaningful stories: Academic advisors often manage hundreds of students with scattered records across multiple systems. AI can aggregate these signals—grades, advising notes, life challenges, and historical context—to produce a coherent narrative that helps advisors respond with empathy and better guidance.

Giving product managers control over prompts is powerful: When PMs own the prompting strategy instead of engineers, they gain direct influence over how AI interprets data and solves user problems. This shifts AI development closer to product thinking—where the focus is storytelling, user pain, and the outcomes the system should prioritise.

Prototyping AI products with synthetic data accelerates innovation: Using generated datasets allows teams to experiment safely, test hypotheses, and validate whether AI can detect meaningful signals. It also enables colleagues to explore prompts, break the system, and collaboratively refine how AI behaves.

AI adoption inside organisations often starts with one brave experiment: Many teams are still figuring out how to work with AI. Jumping into a messy, ambiguous project—despite uncertainty—can rapidly build credibility and create momentum for wider AI adoption across the company.

THE SHOW

Weekly conversations with the AI’s top product leaders. Join Polly Allen as she discovers th

For more on building AI products and careers, along with early course announcement and special pricing, subscribe to the AI Career Boost mailing list at https://aicareerboost.com/interested

THE GUEST

Phil Fairbrother is a product leader working at the intersection of AI, experimentation, and human-centered design. With experience spanning insurance, e-commerce, and creative technology, Phil has built a reputation for turning ambiguity into measurable growth. From launching high-performing chat-based sales funnels at SelectQuote to developing custom CMS platforms that empower independent creators, he thrives on solving complex problems with clarity and momentum. Known for his cross-functional leadership style, Phil blends agile execution with discovery-driven strategy—aligning teams around insight, experimentation, and real user needs. He’s deeply passionate about accessibility, ethical design, and harnessing AI not just to optimize products, but to elevate how teams collaborate and how users experience technology. And in this episode, we’re diving into how AI-powered experimentation, thoughtful design, and product leadership can drive meaningful growth in an increasingly complex digital world.

THE SUMMARY

AI Turns Product Managers Into Founders: You don’t need a technical co-founder anymore to build. With vibe coding and tools like Claude Code, you can go from idea to working product in hours. The barrier isn’t skill — it’s starting.

Agentic AI Is the Real 10x Multiplier: Senior developers using AI as a co-worker can massively increase output — not by blindly accepting code, but by reviewing and directing it. The future isn’t AI replacing devs. It’s AI amplifying the best ones.

Specialised AI Agents > One Big Copilot: Instead of one generic assistant, imagine a team: business analyst agent, brainstorming agent, PM agent writing PRDs. Product workflows can now be systematised and accelerated — especially for greenfield projects.

AI in High-Trust Industries Requires Restraint: In regulated spaces like insurance, hallucinations are unacceptable. The smart play isn’t flashy AI — it’s practical use cases like fallback IVRs during peak season or AI sales training.

Vibe Coding Isn’t Enough — You Need Taste: AI tools default to generic design. If you want standout products, you must be explicit about aesthetic, brand and feel. Prompting isn’t technical — it’s creative direction.

“Time to Wow” Is Everything: Modern users expect instant magic. If your AI tool doesn’t prove value in 30–60 seconds, they’ll assume they can do it in ChatGPT themselves.

Build First. Validate Fast. Don’t Overbuild: Just because you can build features instantly doesn’t mean you should. Tokens and time still matter. Bounce ideas off real users before you go deep.

AI Makes Niche SaaS Possible Again: A CMS tailored for indie authors and alternative music publications? That’s viable now. AI reduces build cost so niche markets become profitable opportunities.

Imposter Syndrome Doesn’t Go Away: Even experienced leaders feel like frauds in AI because it’s so accessible. Accessibility doesn’t invalidate expertise. If you’re building and leading — you’re legit.

The Only Advice That Matters: Start Now - Courses, podcasts, experiments — it doesn’t matter how. Waiting is the only guaranteed way to fall behind.

THE SHOW

Weekly conversations with the AI’s top product leaders. Join Polly Allen as she discovers the paths to success in the world of AI.

THE LINKS

Have a question you want us to answer? Send it through to support@aicareerboost.com

Phil Fairbrother

LinkedIn: https://www.linkedin.com/in/phillip-fairbrother

My links