5d ago

Scaling Delivery in Conversational AI: Beyond Demos to Real-World Systems | With Divyanshu Jain

In this episode of the CDI Expert Series: Unprompted, we sit down with Divyanshu Jain, an enterprise AI and product leader with experience scaling conversational and generative AI at global organizations including PepsiCo and Victoria’s Secret & Co. The episode explores what it takes to move from isolated AI projects to sustainable conversational AI delivery at scale. The conversation traces Divyanshu’s journey from hands-on product delivery to building and leading AI practices inside large, complex organizations. They discuss why many conversational AI initiatives fail to gain adoption, and why fundamentals like usefulness, relevance, and accessibility still matter more than hype, tooling, or trend-driven experimentation. The episode highlights practical strategies for leadership, stakeholder alignment, and product-driven approaches to enterprise AI. The episode also covers: Building conversational AI as a product practice, not a one-off projectScaling delivery from early pilots to enterprise-wide adoptionLeading AI teams amid organisational politics and competing prioritiesBalancing experience design, legality, and risk in customer-facing AIEvaluating generative AI maturity beyond demos and proof-of-conceptsThe conversation highlights how enterprise teams can scale conversational AI responsibly, avoid common traps, and build systems that people actually use. So after listening, the question becomes: how do you scale conversational AI responsibly in enterprise environments - building systems that people actually use while balancing innovation, risk, and organizational realities? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Divyanshu Jain https://www.linkedin.com/in/divyanshu-jain-ai-leader/AI & ML Leader | Generative AI, Responsible AI, & Enterprise Architecture

1h 9m

Jun 2

Growing a Conversation Design Practice: The Invisible Work Behind Successful AI | With Grace Hughes

In this episode of the CDI Expert Series: Unprompted, we sit down with Grace Hughes, Conversation Design Director at Accenture Song, to explore what it really takes to grow, scale, and sustain a conversation design practice inside complex organizations. Drawing on her experience across multidisciplinary teams, Grace shares insights into how the role has evolved from scripting flows to shaping product behaviour in AI-driven systems. The episode highlights how large language models are changing the work of conversation designers, from controlling dialogue to defining principles, guardrails, and trust across non-deterministic interactions. The conversation also highlights the systems thinking, collaboration, and governance required to maintain quality, consistency, and user confidence at scale, while keeping conversation design human-centered and resilient in enterprise environments. This episode explores: How the conversation designer role has changed in the age of LLMsBuilding and operating a Conversation Design Center of ExcellenceMaintaining quality and consistency at scaleDesigning for trust, transparency, and user confidenceThe realities of shipping conversational AI in complex enterprise environmentsThe often invisible work behind successful conversational productsPrinciples for scaling practices while preserving human-centered designThis episode is about how conversation design is maturing as a discipline and what it takes to build a practice that can scale without losing its soul. So after listening, the question becomes: how do you grow a conversation design practice that scales effectively, remains resilient, and builds trust - without losing its human-centered soul? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Grace Hughes Conversation Design Director https://www.linkedin.com/in/grace-hughes-a13a591b/

1h 19m

May 26

From Play-Doh Models to Real Systems: Designing Trustworthy Conversational AI | With Nandini Stocker

Designing Conversational AI That is Trustworthy and Empathetic | From Scripts to System Behaviour In this episode of the CDI Expert Series: Unprompted, we sit down with Nandini Stocker, a global leader in conversational AI with over 25 years of experience designing voice and chat systems across 60+ languages and 120 countries. Drawing on her work at Google, Flipkart, and USAA, Nandini shares a powerful perspective on how conversation design is fundamentally changing in the era of large language models. Together, they explore the shift from rule-based and intent-driven systems to generative, adaptive models - captured in Nandini’s “Legos to Play-Doh” analogy. The conversation focuses on what this transformation means for designers today: moving beyond writing dialogue to shaping system behaviour through tone, guardrails, and governance. They also examine how empathy must now be designed at scale, accounting for emotional context, cultural nuance, and real-world complexity across global audiences. This episode explores: The shift from scripted dialogue to generative, LLM-driven systemsDesigning system behaviour instead of individual promptsBuilding empathy at scale through rules of engagementBalancing global standards with deeply local, context-specific designDesigning for multilingual and culturally diverse marketsGovernance, tone systems, and safety guardrails in high-trust domainsEmpowering designers through tooling, rituals, and shared ownershipThis highlights how design is evolving beyond scripts and prompts, and what it truly takes to build trustworthy, human-centered AI at scale. So after listening, the question becomes: how do you design conversational AI systems that balance flexibility with control, while remaining empathetic, culturally aware, and trustworthy at scale? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Nandini Stocker, Sr. https://www.linkedin.com/in/nandini-stocker Design executive and conversation systems leader with over 25 years of experience shaping how AI speaks across 60+ languages, 120 countries, and 50+ industries.

1h 30m

May 19

Designing Hybrid Conversational AI: Why Hybrid Beats Fully Generative | With Rebecca Evanhoe

Designing Hybrid Conversational AI In this episode of the CDI Expert Series: Unprompted, we sit down with Rebecca Evanhoe to explore what it really takes to design and scale conversational AI in production. Drawing on her work building voice AI for thousands of restaurants at Slang.ai, Rebecca shares a grounded, experience-driven perspective on hybrid systems, evaluation, and why strong conversation design principles matter more than ever in the age of LLMs. We delve into the practical realities of deploying conversational AI at scale, moving past the hype to emphasize that hybrid architectures, combining deterministic logic with carefully scoped LLM tasks, are essential for voice-based and high-stakes environments. A major focus is on rigorous evaluation, where Rebecca outlines the process of designing experiments, defining clear success criteria, and combining human and automated testing to build confidence and mitigate hallucinations, rather than trusting LLMs out of the box. This episode also covers: Why hybrid conversational systems outperform fully generative approachesHow to evaluate LLMs for real-world conversational use casesDesigning voice AI that respects turn-taking, timing, and social normsUsing LLMs as components rather than full conversation driversThe growing importance of evaluation as a core design skillWhat conversation designers should focus on as AI tooling evolvesBalancing automation, hospitality, and brand experience in voice interfacesThis episode offers a practical, no-nonsense look at what it takes to design conversational AI that actually works. So after listening, the question becomes: how do you design conversational AI systems that are not just powerful but reliable, testable, and ready for the real world? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Rebecca Evanhoehttps://www.linkedin.com/in/rebecca-evanhoeAuthor of Conversations with Things, Product Manager at Slang.ai

59 min

May 12

Career Building in Conversational AI: Portfolios, Positioning, and Progression | With Hillary Black

Career Building in Conversational AI: From Scripts to Systems In this episode of the CDI Expert Series: Unprompted, we sit down with Hillary Black for a deep dive into what it really means to build a career and a practice inside the rapidly evolving world of Conversational AI. As the sole conversation designer at Mav, Hillary shares how her role has grown from writing Messenger scripts to shaping product conversations, prompt systems, and design governance across an increasingly complex AI platform. Hillary discusses the realities of startup design: collaborating with engineering and product, transitioning from scripted to prompt-based systems, and maintaining user empathy. She details challenges in versioning, modular design, and deciding the balance between generative AI and human control. This episode covers:- The evolution of Hillary’s role and how she navigates the intersection of design, product, and engineering- Designing complex real-world transitions between messaging, voice, and human handoff- Moving from scripted flows to prompt-based systems - Why empathy and user protection remain core to conversation design- Portfolio dos and don’ts, including why clarity beats volume- Tailoring resumes, identifying transferable skills, and crafting a compelling career narrative- Building a meaningful LinkedIn presence without being “always online”- Why flexibility is becoming the defining skill for conversation designers This episode offers an honest and actionable look at what it means to grow a career in a discipline that’s being rewritten in real time. So after listening, the question becomes: how do you build a career that evolves alongside the technology, while staying grounded in strong design principles and human needs?Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Hillary Blackhttps://www.linkedin.com/in/hillaryblackConversation Designer at MAV

54 min

May 5

Ethics & Ethnography in Conversational AI: Teaching Machines to Listen | With Dr. Liz Rodwell

Designing Human-Centered Conversational AI | Ethics, Education, and EthnographyIn this episode of the CDI Expert Series: Unprompted, we sit down with Dr. Liz Rodwell, anthropologist, UX researcher, and Associate Professor of Digital Media at the University of Houston, to explore how education, ethics, and ethnography shape the way we design and teach conversational AI. Liz brings years of research across classrooms, labs, and cross-cultural fieldwork to show why human insight still matters in an era of large language models. We talk about how teaching design mirrors teaching AI, what cultural awareness means for globally deployed systems, and why the future of conversation design depends on thoughtful, critical, human-centered practices. This episode also covers:- How LLMs are changing the skillset conversation designers need today- Why empathy and communication are still essential- Ethnographic research for uncovering cultural blind spots in AI systems- How educators and learners can help AI better understand nuance, context, and intent- The evolving relationship between design craft and model behavior- How to stay grounded in ethics while working in fast-moving AI environments- Why teaching AI to “listen” starts with teaching humans to notice more- Practical ways to bring cultural/linguistic sensitivity into real-world AI This episode offers a grounded look at how culture, context, and human insight can shape the next generation of conversational systems. So after listening, the question becomes: how do we design and teach AI systems that are not only intelligent, but deeply aware of the human contexts they operate within? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Prof. Elizabeth (Liz) Rodwell, https://www.linkedin.com/in/lizrodwell Media Anthropologist, Educator, Director of UX Lab at University of Houston

1h 5m

Apr 28

Designing Multilingual Conversational AI: Building Globally Aware AI Systems | With Greg Bennett

Designing Multilingual Conversational AI: Prompt Engineering In this episode of the CDI Expert Series: Unprompted, we sit down with Greg Bennett, Head of Prompt Engineering at ZEALS, to explore the evolution from conversation design to prompt design. Drawing on nearly a decade leading AI language strategy at Salesforce, Greg shares a unique perspective on how the field is shifting, from crafting scripted interactions to orchestrating intelligent, agentic systems powered by LLMs.We dive into the role of “taste” in AI design, the balance between creativity and scientific rigor, and the importance of building culturally aware systems in a multilingual, global context. Greg also reflects on his experience working with AI in Japan, highlighting how cultural nuance, context, and adaptability are becoming essential skills for designers shaping the next generation of AI experiences. This episode explores: The evolving journey from conversation designer to prompt designerThe tension between declarative and generative AILessons from multilingual AI development and why cultural context mattersDeveloping intuition and a high bar for qualityBalancing scientific rigor with creativity in AI system designBuilding and managing LLM-powered systems responsiblyEmerging approaches to AI regulation and cultural adaptationThe skills and mindsets designers need to stay relevant in an AI-driven futureThis episode offers a deeply human perspective on what it means to design and lead in the age of intelligent language models. So after listening, the question becomes: how do you design AI systems that balance structure and creativity, while remaining culturally aware, responsible, and impactful? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Greg Bennett Head of Prompt Engineering at ZEALS (Former Director of Conversation Design at Salesforce)https://jp.linkedin.com/in/gab45

1h 7m

Apr 21

Designing Multimodal AI Agents: Inside Google DeepMind’s Gemini Live | With Karen Kaushansky

Designing for Conversation in the Age of Multimodal AI In this episode of the CDI Expert Series: Unprompted, we sit down with Karen Kuczynski, UX lead for Gemini Live at Google DeepMind, to dive into the evolution of design from deterministic flowcharts to open-ended LLMs, the mechanics of building trust in AI, and the role of multimodality in shaping natural, human-like interactions. Karen reflects on her career arc, the resilience it takes to work at the cutting edge, and how designers today must rethink what it means to guide a “conversation” between people and AI systems. This episode explores: How conversational AI is shifting from voice-only to multimodalDesigning Gemini Live, a real-time companion that can see what you see, translate across languages, and collaborate in the momentThe challenge of creating trust frameworks for LLMsWhy voice identity and diversity matter when users form ongoing relationships with AIThe shift from deterministic flowcharts to principles, guidelines, and facilitation between humans and AI modelsThe role of user research and feedback loops in validating hypotheses and iterating productsLessons from designing communication systems for autonomous vehicles and emerging interfacesThis conversation offers a vital perspective, whether you're a veteran in the field or just starting out, on where conversation design has been—and its trajectory in this new era of generative, multimodal AI. The ultimate question after listening is: how do we design conversational experiences that are not only intelligent but genuinely human-centered, adaptive, and trustworthy? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Karen Kaushansky Conversational AI/UX Lead at Google https://fr.linkedin.com/in/karenkaushansky

1h 2m

Apr 14

Building Predictable Conversational AI: Design, Engineering & Control | With Yam Marcovitz

Building Predictable Conversational AI: Bridging Design, Engineering & Control at Scale In this episode of the CDI Expert Series: Unprompted, we sit down with Yam Marcovitz, former software architect at Microsoft and Apex.AI, and co-founder of the open-source AI framework Parlant. With a background in real-time systems and enterprise software, Yam brings a systems-level perspective to one of AI's biggest challenges: getting LLMs to consistently do what you want.The conversation explores the gap between design and engineering in conversational AI, why current tools fall short, and how open source plays a crucial role in responsible AI development. Yam shares practical insights on building agent behavior that's not just intelligent but reliable and predictable at scale. The episode also explores: How team collaboration and clear communication impact AI behavior as much as the technology itselfThe limitations of current conversation design toolsSeparating design and engineering workflows to improve efficiency and clarityCreating flexible agents without sacrificing controlMoving beyond "prompt and pray" to systematic testing and validationManaging updates/fixing issues without retraining models from scratchDesigning for multilingual, multichannel interactionsHow linguistic precision directly impacts agent performanceWhy open source and transparency are key to building responsible, scalable AI systemsThroughout the episode, Yam offers a practical, behind-the-scenes look at how robust AI systems are actually built; balancing flexibility with control, and creativity with structure. So, after reflecting on this episode: how can you build AI systems that are not just intelligent, but truly reliable, testable, and scalable? Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Yam Marcovitz,https://il.linkedin.com/in/yam-marcovicCEO, Parlant.io

33 min

Apr 7

Optimizing AI: Strategy, Collaboration & Real-World Impact | With Lauren Goerz

Optimizing Conversational AI in the Enterprise: Strategy, Collaboration & Real-World Impact In this episode of the CDI Expert Series: Unprompted, we sit down with Lauren Goerz, Technical Marketing Engineer at Rasa, who brings nearly a decade of experience designing and deploying conversational AI across industries. Lauren walks us through what it really takes to build systems that don't just work, but actually get better over time. The conversation centres on the importance of organisational alignment, highlighting how successful AI adoption depends on collaboration across leadership, product, engineering, and customer teams. Lauren also discusses the growing role of conversation designers in shaping experiences that are not just functional, but intuitive and human. The episode also explores: Why one-size-fits-all doesn't work in conversational designConversation review as a core disciplineProving value in the enterprise: How to align with stakeholders, track metrics that actually matter (like topic-level success rates)The real impact of AI on costs, satisfaction, and service quality.The case for hybrid systems: Rasa's "process calling" approach, a structured way to blend LLMs with deterministic flows for safety, scalability, and control.How teams can build systems ready for long-term iteration using metadata, user feedback, and strategic measurement as their foundation.What modern Conversational AI teams actually look like Whether you're wrestling with enterprise AI challenges or just trying to make your assistant a little better each week, this conversation offers both big-picture clarity and actionable insights you can use right away. So, after reflecting on this episode: how can you better align teams, tools, and strategy to scale conversational AI effectively in your organisation? - Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Lauren Goerz https://www.linkedin.com/in/lauren-goerz/ Senior Technical Marketing Engineer at Rasa

57 min

Mar 31

Persona in Conversational AI: Designing with Empathy, Trust, & Ethical Care | With Jason Gilbert

Persona in Conversational AI: Designing AI Companions with Empathy, Trust, & Ethical Care In this episode of the CDI Expert Series: Unprompted, host Michele sits down with Jason Gilbert, Lead Character Designer at Intuition Robotics, to explore the thoughtful design behind Lark, a companion robot created to support older adults who are aging in place. Drawing on his experience, Jason shares how Lark was built to strengthen human relationships, not replace them. He highlights the unique challenges of designing technology for a vulnerable population. Jason explains that trust, transparency, and user autonomy are central to Lark’s design, from safeguarding privacy to ensuring users remain in control. The episode also explores: How conversational AI can support older adults without replacing human relationshipsThe role of multimodal design (voice, lights, gestures, and screen) in creating more natural interactionsUsing generative AI to encourage creativity through art, poetry, and storytellingThe ethical responsibilities of designing empathetic AI systems Throughout the discussion, Jason shares insights into the interdisciplinary work behind conversational AI, highlighting collaboration between designers, engineers, and psychologists to build technology that is both helpful and emotionally intelligent. The episode offers valuable perspectives for anyone interested in human-centred conversational AI design and the future of compassionate technology. So, after reflecting on this episode: how can designers create technology that supports human connection rather than replacing it?- Contact Us Thanks for tuning in to Unprompted, for more practical lessons, and behind-the-scenes stories, exclusive events, and candid insights, visit our website or follow us on your favorite social platform. Follow us on social media:X (formerly Twitter): https://x.com/CDInstitute_Instagram: https://www.instagram.com/Conversation_Design_Institute/Facebook: https://www.facebook.com/ConversationDesignInstitute/ Find out more:Website: https://www.conversationdesigninstitute.com Connect with us:LinkedIn: https://www.linkedin.com/company/conversation-design-institute Jason GilbertLead Character Designer at Intuition Roboticshttps://www.linkedin.com/in/jason-f-gilbert-75186a156

48 min

Unprompted

Episodes

Scaling Delivery in Conversational AI: Beyond Demos to Real-World Systems | With Divyanshu Jain

Growing a Conversation Design Practice: The Invisible Work Behind Successful AI | With Grace Hughes

From Play-Doh Models to Real Systems: Designing Trustworthy Conversational AI | With Nandini Stocker

Designing Hybrid Conversational AI: Why Hybrid Beats Fully Generative | With Rebecca Evanhoe

Career Building in Conversational AI: Portfolios, Positioning, and Progression | With Hillary Black

Ethics & Ethnography in Conversational AI: Teaching Machines to Listen | With Dr. Liz Rodwell

Designing Multilingual Conversational AI: Building Globally Aware AI Systems | With Greg Bennett

Designing Multimodal AI Agents: Inside Google DeepMind’s Gemini Live | With Karen Kaushansky

Building Predictable Conversational AI: Design, Engineering & Control | With Yam Marcovitz

Optimizing AI: Strategy, Collaboration & Real-World Impact | With Lauren Goerz

Persona in Conversational AI: Designing with Empathy, Trust, & Ethical Care | With Jason Gilbert

About

Information