Interview #69 Michael Wu, Chief AI Scientist at PROS

Dr. Michael Wu, Chief AI Scientist at PROS, discusses his transition from computational neuroscience research to applied AI in industry, emphasizing how his perspective without domain-specific baggage has enabled innovative problem-solving approaches. He explores the evolving landscape of AI agents and tools, highlighting the importance of Model Control Protocol (MCP) as a bridge between passive language models and actionable AI systems that can interact with existing enterprise tools. Wu emphasizes the need for organizations to create fail-safe environments that encourage AI experimentation while maintaining security, and advocates for balancing innovation speed with responsible development practices that prioritize safety, privacy, and legal compliance from the outset.

Maddie Daianu, Head of Data & AI at Intuit Credit Karma, focuses on building sophisticated recommendation systems that personalize financial offerings for over 100 million members by creating unified consumer profiles that track complete financial journeys. She emphasizes the critical importance of maintaining strict compliance requirements in the FinTech industry while leveraging both traditional machine learning for core systems and selective use of generative AI for contextualization. Her approach includes rigorous evaluation frameworks with five key metrics—product alignment, safety, compliance, data accuracy, and system integration—ensuring all AI implementations meet the high standards required for financial services.

Eleanor Lightbody, CEO of Luminance, explains how generative AI transformed legal work by providing 10x improvements over traditional machine learning systems, enabling end-to-end contract automation rather than just document review. She emphasizes that trustworthy AI in legal contexts requires multiple models checking each other's work to ensure consensus, with systems reverting to human oversight when confidence thresholds aren't met, particularly crucial for high-stakes deals and regulatory compliance. Lightbody discusses how AI agents will evolve beyond assistants to proactively complete work and seek approval, fundamentally changing how professionals interact with software and potentially disrupting traditional workflows while creating new opportunities.

Join Krish Ramineni, CEO of Fireflies.ai, as he explores the transformative impact of AI on workplace productivity and the future of work. Ramineni discusses how the convergence of improved transcription technology, remote work adoption, and large language models has revolutionized how we approach meetings and collaboration. He shares insights on the evolution from basic speech-to-text to intelligent AI teammates that can understand context, provide real-time assistance, and automate routine workflows. The discussion covers practical strategies for leveraging AI tools across different industries, the importance of learning to prompt effectively, and predictions about how small teams will build billion-dollar businesses with AI assistance. Ramineni also offers advice on embracing AI tools for content creation, research, coding, and customer support, emphasizing that speed and adaptability will be crucial differentiators in the AI-powered workplace of the future.

Join Tanmai Gopal, CEO at Hasura, as he discusses the current state of RAG and AI tooling challenges. Gopal explains that RAG has cooled down from its peak popularity, noting that while it works for low-hanging fruit use cases, it fails in mission-critical scenarios where good enough is not the threshold and precision is required. He describes RAG's fundamental limitation as not understanding context, comparing it to using a single hammer for everything, and advocates for agentic approaches that combine RAG with other tools. Gopal predicts that AI reliability will become the dominant topic by year's end, as he observes that newer models have become skilled at convincing themselves and users that they're correct when they're actually hallucinating.

Join Stefan Klopp, CTO at Matador Network and creator of Guide Geek, as he explores how artificial intelligence is revolutionizing the travel industry. Stefan discusses the shift from traditional video-focused destination marketing to AI-powered trip planning. He reveals how Guide Geek evolved from a consumer chatbot on WhatsApp and Instagram into a B2B SaaS platform helping destination marketing organizations worldwide. Stefan addresses the challenges of AI-generated fake reviews, the anxiety-reducing benefits of having AI travel assistants in your pocket, and why destinations are eager to adopt AI despite concerns about hallucinations and incorrect information. He predicts that agentic AI will be the next breakthrough, creating autonomous travel concierges that can plan, book, and manage entire trips seamlessly.

Join Gavin Myers, Managing Partner at Prudence VC, as he discusses how artificial intelligence is transforming real estate, construction, and infrastructure investing. Gavin shares insights on what successful founders need to demonstrate and explains how AI adoption differs from previous tech cycles, with unprecedented top-down enterprise demand driving faster growth for startups. He explores practical AI applications in commercial real estate and construction, from automating mortgage processing to using drone imagery for construction site monitoring. Gavin predicts the next frontier will be multi-agent architectures handling tedious tasks like preparing offering memorandums, buyer outreach, and property sourcing—transforming industries facing workforce shortages and increasing investment.

Join Babak Hodjat, CTO of AI at Cognizant, as he discusses two major AI developments. Hodjat explains how DeepSeek, an open-source model from China, disrupted the AI landscape with unprecedented efficiency—running 14B parameter models on consumer hardware while achieving performance comparable to much larger models, which temporarily impacted AI-related stocks. We then explore multi-agent systems, which transform large language models into agents by enabling them to use tools and execute actions in the world rather than just generating text. These agent systems can coordinate with each other in natural language, operate autonomously within defined boundaries, and collaborate to solve complex tasks while maintaining human oversight.

Join Ron Green, CTO of KUNGFU.AI, as he discusses the evolution of artificial intelligence from its early underestimated beginnings to today's generative AI boom sparked by ChatGPT. Ron shares insights on common misconceptions about AI implementation, emphasizing the importance of clean first-party data and the need for human oversight when deploying AI solutions. He also highlights how proprietary data is the key to valuable AI investments, while cautioning against rushing implementation without proper safeguards.

Join Barr Moses, CEO of Monte Carlo, as she discusses the crucial role of data and AI observability in building reliable AI products. She explains how enterprises can gain competitive advantage by leveraging their first-party data and implementing proper data quality monitoring systems. Moses highlights that regardless of industry, organizations face similar challenges with data reliability, which can be traced to four root causes: problems with the data itself, code issues, system failures, and model output errors.