Winners' Circle

Jim Napolitano on 3DLive, Virtual Twins, and the Future of Immersive Product Design

Jim Napolitano is helping brands, engineers, product teams, and enterprise leaders experience product development in a more immersive way. As North American Services Director at Dassault Systèmes 3DEXCITE, Jim works with teams using virtual twins, spatial computing, and the 3DLive application to bring complex product data into interactive 3D environments. Dassault Systèmes recently won an innovation award for 3DLive and its work with Apple Vision Pro. In this episode, Russ and Jim explore how virtual twins are changing the way companies design, simulate, manufacture, market, and improve products. Jim explains how a virtual twin becomes a data fueled version of a product, allowing teams to test, review, and understand products before they exist physically. They dive into 3DLive and how immersive technology helps engineers, executives, designers, marketers, and end users collaborate inside product data. Jim shares examples from Formula 1, automotive design, aviation, consumer packaging, manufacturing training, and even medical applications like a living heart. The conversation also covers why live data matters, how spatial computing can make product review more intuitive, and how partnerships with Apple and Nvidia are helping Dassault Systèmes bring virtual twin experiences to life in new ways. Along the way, Jim discusses immersive collaboration, faster decision making, physical prototyping, simulation, AI, training scenarios, consumer research, and why virtual twins may soon become central to how companies build and improve nearly everything. Topics Covered: [00:01] Welcome and intro, Jim Napolitano and Dassault Systèmes’ 3DLive award win [00:39] What Dassault Systèmes does and the role of virtual twins [01:25] Connecting agency leadership, brand work, and immersive technology [02:41] What a virtual twin is and how it differs from an agent [04:02] Spatial computing, sense computing, and intuitive product interaction [05:14] Using Formula 1 to show simulation data in immersive 3D [06:30] Exploring vehicle components inside a virtual twin [07:32] How 3DLive supports collaboration across different locations [09:00] Making immersive tools useful for non technical users [10:00] Using 3DLive for consumer packaging research [11:07] Replacing expensive physical iteration with immersive product review [11:50] Aviation use cases and reconfiguring passenger cabin spaces [13:20] What surprised Dassault Systèmes during real customer deployments [14:00] Collaboration with Apple on Apple Vision Pro experiences [15:26] Knowing when a virtual twin is good enough for decision making [16:13] Using existing product data instead of creating one off experiences [17:10] Training technicians with virtual equipment and troubleshooting scenarios [18:22] Why live data matters in product development [20:05] Dassault Systèmes’ partnerships with Apple and Nvidia [21:25] Common principles across automotive, aerospace, and medical use cases [22:00] The living heart example and virtual twin applications in medicine [24:28] How immersive tools and AI can shorten product development cycles [25:17] How AI may support engineering, science, and strategic decision making [26:27] Final thoughts on the future of 3DLive and virtual twin technology

Alexey Sheremetyev on Turning Your Home Into an Editable Digital Twin

Alexey Sheremetyev is helping homeowners, designers, real estate professionals, builders, and contractors reimagine how physical spaces become digital. As Co-Founder and CPO of Planner 5D, Alexey helped build a platform that allows users to scan rooms with a phone camera and turn them into editable 3D home plans. Planner 5D recently won an AI Excellence Award for its Home Scanner technology. In this episode, Russ and Alexey explore how Planner 5D grew from a personal renovation problem into a platform used by more than 100 million people around the world. Alexey shares how his background in web design and user experience shaped the product, and why the goal was always to make home planning simple enough for consumers but powerful enough for professionals. They dive into Home Scanner, Planner 5D’s AI powered feature that uses computer vision to turn real rooms into editable digital twins. Alexey explains how the technology can recognize room layouts, measurements, furniture, colors, textures, flooring, and objects without requiring expensive hardware or specialized training. The conversation also covers why editable 3D plans matter more than static renderings, how AI helps handle messy real world spaces, and why Planner 5D’s years of user generated floor plans and designs have become one of its most valuable assets. Along the way, Alexey discusses renovation planning, real estate workflows, professional design collaboration, smart home possibilities, home maintenance, property history, and his vision for Planner 5D becoming a persistent digital memory for the home itself. Topics Covered: [00:01] Welcome and intro, Alexey Sheremetyev and Planner 5D’s AI Excellence Award win [00:29] What Planner 5D does for homeowners and professionals [00:37] Creating digital twins and editable 3D home plans [01:27] Turning room design into a consumer grade 3D experience [01:56] How Alexey’s own apartment renovation inspired Planner 5D [03:53] Why traditional room planning can be frustrating and inaccurate [04:18] How a design background shaped the Home Scanner experience [05:00] Using AI to automate manual measurements and room recreation [05:59] Why Home Scanner is technically harder than it looks [06:17] Planner 5D’s advantage after 15 years and more than 100 million users [07:35] Solving spatial reconstruction through software instead of costly hardware [07:56] Using computer vision instead of relying only on LiDAR [10:12] What Planner 5D learned from real users scanning real spaces [11:00] How real estate professionals use digital twins in their workflows [12:14] Building for both consumers and professional users [13:20] How builders and contractors use Planner 5D as a presale tool [14:27] Why editable 3D plans create a different user experience than static images [15:50] How Planner 5D checks scan accuracy for renovation planning [16:18] AI validation, human review, user feedback, and correction tools [18:51] Handling furniture, bad lighting, unusual room shapes, and messy spaces [19:30] Training AI with a large base of floor plans, designs, and user data [22:00] Using Planner 5D as a filing cabinet and memory system for the home [24:10] Why spatial reconstruction may go beyond home design [25:57] Planner 5D as the next stage of the smart home [27:46] Why the digital twin could travel with the house, not the owner [28:00] Smart home integrations, appliances, and connected home systems [29:30] Helping homeowners understand and maintain older properties [30:00] Using AI to recommend repairs and improvements that may increase home value [32:00] Why user expectations for instant answers and context are changing [33:59] Final thoughts on the future of Planner 5D and generative AI

Anton Dam on Betting on Human Creativity in AI, Risk, and Internal Audit

Anton Dam is helping enterprise risk, audit, and compliance teams rethink how AI can support highly regulated work without removing human judgment from the process. As SVP of Product and AI at Optro, formerly AuditBoard, Anton works with global enterprises using AI to make risk management more adaptable, strategic, and aligned with business goals. Optro recently won an AI Excellence Award for its work bringing AI into audit, risk, and compliance workflows. In this episode, Russ and Anton explore how internal audit and GRC teams are using AI to move beyond manual coordination, document review, and repetitive evidence analysis. Anton explains why audit teams often spend the majority of their time on labor intensive review work, and how AI can help shift that effort toward higher value risk management. They dive into Optro’s approach to assistive AI, copilot experiences, and agentic workflows. Anton shares why quality thresholds change depending on the use case, why human review remains a nonnegotiable, and why nothing should enter a system of record without a person in control. The conversation also covers AI governance, shadow AI, regulatory change, customer trust, and why large, highly regulated companies may be moving faster on AI adoption than many people expect. Anton explains why Optro does not train on customer data, how the company thinks about AI transparency, and why governance will become as central to business operations as cybersecurity. Along the way, Anton discusses human creativity, AI alignment, enterprise trust, AI security reviews, audit team adoption, the future of web apps, and why internal audit may become a more strategic advisory function as AI takes on more repetitive work. Topics Covered: [00:01] Welcome and intro, Anton Dam and Optro’s AI Excellence Award win [00:27] Optro’s mission in audit, risk, and compliance [01:09] Why GRC is changing in a more regulated AI environment [01:28] Anton’s path from LinkedIn and Workday to Optro [04:41] Why Anton believes in betting on human creativity [05:00] What AI can automate, and what remains deeply human [08:16] A typical day for internal audit teams before AI [09:00] Manual coordination, evidence gathering, and document review [10:00] How AI can reduce time spent on repetitive audit tasks [12:00] What good enough AI means in high stakes risk and compliance work [12:26] Assistive AI, copilot workflows, and agentic AI [14:48] Why human review remains required before records are updated [15:29] Staying current with changing regulations and standards [16:24] Tracking data sets, model tuning, and development decisions [17:44] Building trust with enterprise customers [18:12] Why quality and workflow fit drive AI adoption [20:03] What surprised Anton about AI adoption in large enterprises [22:57] AI security reviews and integrating into enterprise AI ecosystems [23:20] Why the traditional web app may change dramatically [24:24] Measuring AI impact in risk management [26:23] Optro’s nonnegotiables for deploying AI [26:35] Why Optro does not train on customer data [27:21] Using AI governance to help organizations govern AI [28:30] Why AI governance may follow the same path as the CISO function [31:17] Shadow AI and lack of visibility inside organizations [31:50] The risks of employees using unapproved AI tools [32:30] Why companies must enable AI safely instead of simply blocking it [33:40] What internal audit could look like in five years [34:20] Moving from risk mitigation to risk management [35:12] Final thoughts on internal audit as a strategic advisory function

Securing the Human Data Layer of AI with Siobhan Hanna

Siobhan Hanna is helping AI companies protect one of the most important parts of model development: the human data layer. As a leader at WeLo Data, she works with foundational LLM builders and enterprise technology companies to provide high quality multilingual human data across languages, cultures, and markets. WeLo Data’s NEMO framework recently won an AI Excellence Award for helping detect fraud, misrepresentation, and data integrity risks in AI training pipelines. In this episode, Russ and Siobhan explore why high quality human data is essential to building better AI models, and why that data is increasingly vulnerable to fraud. Siobhan explains how contractor based, globally distributed AI data workflows can create opportunities for identity fraud, coordinated manipulation, account sharing, and other risks that can degrade model performance. They dive into NEMO, WeLo Data’s fraud mitigation and misrepresentation detection framework. Siobhan shares how the system uses continuous monitoring, behavioral analytics, rules based logic, AI driven detection, and organizational psychology to identify suspicious activity across the contributor life cycle. The conversation also covers why AI data integrity should be treated as part of the broader data quality and governance conversation. Siobhan explains why point in time checks are not enough, how WeLo Data borrowed ideas from financial services and KYC models, and why continuous monitoring is critical when training data is so strategically valuable. Along the way, Siobhan discusses multilingual AI, cultural context, data provenance, contributor verification, regulatory trends, and why protecting the human layer of AI development may soon move from best practice to formal requirement. Topics Covered: [00:01] Welcome and intro, Siobhan Hanna and WeLo Data’s AI Excellence Award win [00:28] WeLo Data’s role as a multilingual AI human data provider [01:05] Why AI training data quality matters [01:24] How fraud can enter human data workflows [02:29] Why fraud mitigation in AI data has been underserved [02:36] The speed of AI development and the blind spot around human data integrity [04:28] How fraudulent or misrepresented data can affect model performance [04:57] Why data integrity issues can be hard to trace after model degradation [06:08] Why fraud is difficult to detect in global AI data pipelines [07:02] Which AI systems are most exposed to training data integrity risks [08:10] Identity validation and why AI data fraud differs from traditional fraud [08:35] Borrowing KYC and transaction monitoring ideas from financial services [10:27] How WeLo Data validates that NEMO is catching the right activity [11:24] Behavioral variables, rules based detection, and AI driven monitoring [13:04] The role of organizational psychology in fraud detection [13:53] Stopping threats before they reach the model [14:28] What surprised WeLo Data about the AI fraud landscape [15:30] Why multilingual and cultural context make fraud detection harder [17:02] Why continuous monitoring beats one time screening [18:04] What translated from financial services and what had to be reinvented [19:20] AI regulation, data integrity, and governance requirements [19:48] Why contributor verification may become a formal AI requirement [20:50] Why data provenance should be part of responsible AI infrastructure [21:23] Questions AI companies should ask about who produced their data [22:43] Which parts of AI infrastructure are most vulnerable [23:04] Advice for AI founders, operators, and leaders [23:53] Final thoughts on fraud, trust, and protecting AI training data

AI Agents for Customer Service with Latane Conant

Latane Conant is helping companies rethink customer service as a relationship builder, not just a cost center. As CMO of Parloa, she is working at the intersection of AI agents, voice, customer experience, and enterprise support, helping companies replace outdated IVR systems with conversational AI that can make every customer interaction feel as easy as talking to a friend. In this episode, Russ and Latane explore why the customer service side of the buyer journey has become one of the biggest missed opportunities in business. Latane explains how companies spend heavily to get customers to engage, but often fail them when they actually need help. They dive into Parloa’s AI voice agent platform and how it helps enterprises deliver secure, low latency, natural language conversations across languages, dialects, and customer scenarios. Latane explains why voice is the hardest modality to get right, why reliability matters more than flashy demos, and why regulated industries need AI that can handle authentication, tool calling, context, and secure interactions at scale. The conversation also covers Parloa’s AI mystery shopping study of the Fortune 2000, which found major gaps in customer support access, chat resolution, IVR experiences, and agent readiness. Latane shares why she believes companies need to prepare for an agent to agent future where customers may soon expect their personal AI agents to interact directly with enterprise systems. Along the way, Latane discusses customer journey leaks, the limits of “check the box AI,” the importance of use case selection, enterprise deployment timelines, simulation testing, agent drift, and why customer service should become a driver of loyalty, revenue, and lifetime customer value. Topics Covered: [00:01] Welcome and intro, Latane Conant and Parloa’s award wins [00:21] Why Latane moved from 6sense to Parloa [00:32] The customer journey leak inside customer service [02:09] Why CMOs should care about support and service experiences [02:44] Parloa’s mission to make customer interactions feel like talking to a friend [04:14] How Parloa differs from basic LLM-based call tools [04:43] Replacing outdated IVR systems with conversational AI [06:20] Why traditional IVR experiences lose context and frustrate customers [07:03] Parloa’s AI mystery shopping study of the Fortune 2000 [08:08] Why many companies hide or limit customer support access [08:33] Chatbot resolution rates and poor human handoff performance [09:02] Why only 1% of companies are ready for agent to agent interactions [09:58] The coming wave of personal AI agents contacting enterprises [10:38] AI agents as relationship builders, not just transaction handlers [10:49] Travel, payments, insurance, and roadside assistance use cases [13:11] Solving context loss across customer service interactions [13:41] Building a broader customer context fabric [15:16] Deploying AI agents at enterprise scale [15:38] Parloa’s foundation in real-time translation and voice technology [17:48] Why AI can accelerate customer service deployments [18:17] Fast enterprise deployment through use case prioritization [19:50] Prebuilt integrations and reusable AI skills [20:35] Why AI agents need training before going live [22:48] Reliability, authentication, tool calling, and production latency [24:43] Transactional versus high stakes customer service interactions [26:37] How customer comfort with AI will evolve over time [27:26] Common mistakes executives make when deploying service AI [28:25] Why companies should rethink the front door of customer service [29:12] Customer service as an opportunity to build loyalty [30:09] Final thoughts on personal agents and the future of customer experience