Artificial Reality

LJB Engineering fed a legacy macro into Claude. It cut days of work to minutes. Matt Bell explains what happened next.AI adoption speaker and Artificial Reality host Galina Fendikevich sits down with Matt Bell, Director of Process and Quality Improvement at LJB Engineering, to break down how a firm founded in 1966 is adopting AI across six engineering disciplines without a company-wide mandate or a massive rollout. Matt has spent nearly two decades in structural engineering, project management, and geospatial services. He now leads digital transformation at LJB, a firm designing buildings, bridges, tunnels, fall protection systems, and transportation infrastructure for public and private clients.Matt's approach to AI adoption starts with people, not strategy. He found engineers at LJB paying for AI tools out of their own pocket and spending off-hours learning. Instead of shutting it down, the firm gave those employees dedicated weekly time, kept them embedded in their engineering teams, and let them drive adoption from inside each practice. Galina calls it "pick the headache": find the one task nobody wants to do, the work that is slow and manual, point AI at that single problem, and let the results pull the rest of the team forward.In this episode, Galina and Matt cover where AI is working in civil engineering and where it is not:The concrete wall reinforcing spreadsheet that was costing junior structural engineers an entire day of guess-and-wait iterations. Matt's team fed the legacy VBA macro into Claude, which identified the bottleneck and helped rebuild it. Each design iteration now returns results in seconds instead of minutes.Why a third-party AI quality control tool returned thousands of coordination issues pulled from other firms' standards, almost none relevant to LJB. Processing the false positives took longer than manual review. Matt explains why building internal AI tools trained on your own standards and workflows delivers better results than off-the-shelf platforms.Matt's biggest concern about generative AI in engineering: that junior staff will develop a dependency on AI outputs and lose the critical thinking that only comes from designing structures, walking job sites, and learning directly from experienced contractors.What 60 years of institutional knowledge looks like when it lives entirely in senior employees' memories. Matt describes the search problem LJB faces and what he wishes AI builders would create: a system to catalog decades of engineering drawings, tag them by project type and feature, and make past solutions searchable for current projects.This episode is for you if you work in civil engineering, structural engineering, architecture, construction, or consulting and want to see how a mid-size firm is practically adopting AI. It is also for business leaders who want an adoption model that starts with one frustrating workflow instead of a strategy deck, and for AI builders who want to hear what the people using their tools actually need.Follow Matt Bell:https://www.linkedin.com/in/matthew-bell-p-e-9948526/LJB Engineering: LinkedIn: https://www.linkedin.com/company/ljbinc/ https://www.ljbinc.com/Artificial Reality:artificialrealitymedia.cominstagram.com/artificialrealitymedialinkedin.com/company/artificial-reality-mediagalinafendikevich.com#ArtificialReality #ArtificialRealityPodcast #AIAdoption #CivilEngineering #DigitalTransformation

Net Road Show’s Robb Lillie breaks down how ElevenLabs replaced 3AM banker recording sessions, why Microsoft Copilot drove employees to use unsanctioned AI tools, and the EU AI Act compliance gap he calls “fast car, no seat belt.”Robb Lillie has spent 22 years at Net Road Show, the platform behind investor communications for Goldman Sachs, Morgan Stanley, and Bank of America. He joined when the company was part of broadcast.com before Yahoo’s $5.7 billion acquisition, and now leads AI adoption across text-to-speech, proof-of-concept building, and training for a team that powers IPO roadshow and M&A deal workflows worldwide.In this episode, Galina Fendikevich and Robb cover what’s working, what’s broken, and what still needs to be built in AI adoption for financial services infrastructure:How ElevenLabs text-to-speech eliminated overnight recording sessions for international bankers, including a Korean banker in tears at 3AM trying to get through an English script Why Microsoft Copilot is so weak inside enterprise environments that employees go around IT to use ChatGPT and Claude, creating a shadow AI problem that governance and compliance teams cannot track Robb’s framework for evaluating where AI belongs in enterprise workflows: “Coordination collapses, but the context doesn’t.” AI handles logistics, but judgment and creativity stay human. A wish list for AI builders: unified search across seven enterprise systems, compliance tooling for AI-generated video under the EU AI Act, and a correction memory layer for translated audio How Robb discovered his company’s best AI use cases by polling his team on their biggest pain points instead of asking what AI could do for themThis episode is for anyone working in financial services, capital markets, or regulated industries who wants to see how a 22-year deal event production company is adopting AI tools like ElevenLabs, Lovable, and Synthesia without cutting corners on compliance. It’s also for business leaders evaluating enterprise AI who need a practical framework for deciding what to automate and what to leave alone. And if you’re a builder or developer, Robb names specific unsolved problems in enterprise AI, from search layers across legacy systems to compliance verification for AI-generated video to audio correction memory that learns business terminology.Follow Robb Lillie:LinkedIn: https://www.linkedin.com/in/robblillie/ Website: netroadshow.com Email: [email protected] Reality: Website: artificialrealitymedia.com Instagram: instagram.com/artificialrealitymedia LinkedIn: linkedin.com/company/artificial-reality-media Galina Fendikevich: linkedin.com/in/gfendi

Keynote speaker James Taylor breaks down how he runs his speaking business on Claude, why cheap AI pricing won't last, and how 1,000 people made a Taylor Swift reggae theme song with AI in about 15 minutes.On Artificial Reality, I ask business owners what's working with AI, what's broken, and what still needs to be built. This episode is a behind-the-scenes look at how one of the world's busiest keynote speakers uses AI across sales, marketing, speech writing and delivery.James Taylor is an international keynote speaker on creativity, artificial intelligence and innovation, and the author of SuperCreativity: Augmenting Human Creativity in the Age of Artificial Intelligence. He started his career managing rock stars, moved into tech in Silicon Valley, and now gives 50 to 100 keynotes a year in 25+ countries for clients including Apple, Deloitte, Visa and McDonald's. He is a Fellow of the Royal Society of Arts.In this episode, I ask James what's working, what's broken, and what still needs to be built with AI in his speaking business:● How a keynote speaker uses AI day to day: James has an AI chief of staff built on Claude that briefs him every morning and, 24 hours before a sales call, sends psychometric profiles of everyone joining. A Claude "story bank" of every story, framework and quote he has used drafts the first version of each keynote. He then rewrites it himself.How to get recommended by ChatGPT and Claude: his team builds thousands of highly specific landing pages optimized for AI answer engines, using Claude for research, Perplexity to create pages at scale and Gemini for quality checks. James says chatbots now name him a top AI keynote speaker in Amsterdam, a city he has never lived in.Why cheap AI won't last: James argues today's AI prices are subsidized by venture capital, like early Uber, and that businesses should plan for higher rates by matching cheaper models to simpler tasks. He also sees low-cost Chinese models putting pressure on US AI companies.Is AI replacing jobs: his own team shrank, including a video editor role, while a customer service hire was retrained in SEO, AEO and GEO. James believes efficiency is only phase one of AI. The bigger opportunity is creating things that don't exist yet.What Silicon Valley should build next: James wants an AI coach that reviews presentation videos across 100 factors, plus live translation so non-native English speakers can ask questions. I share why nearly every guest I've interviewed wants better voice AI.This episode is for you if you're a speaker, consultant or creator who wants AI to handle the business side of your expertise, or a business owner figuring out where AI actually saves time and where it still falls short. It's also for founders and developers building AI tools who want to hear what a power user running five AIs at once still can't find on the market.Follow James Taylor:jamestaylor.meyoutube.com/@JamesTaylorMelinkedin.com/in/jameskirktaylorinstagram.com/jamestaylordotmeSuperCreativity: geni.us/QiDBuArtificial Reality:artificialrealitymedia.cominstagram.com/artificialrealitymedialinkedin.com/company/artificial-reality-medialuma.com/artificialrealityHosted by Galina Fendikevich

Steven Webster sold his company to Adobe, consulted for NASA, NATO, and Nike, and built a full AI operating system for his next company. Here's how.Steven Webster is the founder and CEO of Asensei, the movement recognition company that does for physical movement what speech recognition did for voice. Before Asensei, he sold his first company to Adobe, built their worldwide design consulting practice, and opened an innovation studio for Microsoft. He was also the UK's youngest junior jiu-jitsu black belt and coached Edinburgh University's Karate Club to 10 consecutive national championships.In this episode of Artificial Reality, I ask Steven what's working with AI, what's broken, and what still needs to be built for CEOs running real companies:How 5 martial arts principles (Shoshin, Keiko, Mushin, Sensen no sen, and teaching others) became a white belt to black belt AI adoption framework that any business leader can follow step by stepHow Steven built an "agentic office of the CEO" with AI agents running sales, contracts, RevOps, product insights, and a Friday performance coaching session with an agent named Wendy, modeled after Wendy Rhodes from BillionsWhy he compares AI adoption to driving a Model T Ford. Users are still stuck managing context windows and model selection when they shouldn't have to be. His prediction: AI tools will go from manual to automaticThe moment he caught himself saying "you're using it wrong" and stopped. His design thinking background at Adobe taught him to never blame the user, always fix the productHis vision for "multiplayer mode" where AI agents join meetings as participants, teams share versioned skills across departments, and memory travels with you across different AI models instead of being locked inside oneThis episode is for you if you're a CEO or founder looking for a concrete AI adoption framework from someone who already built one, if you want to understand how to go from basic AI chat to a full agentic operating system across your company, or if you're a builder trying to understand why business leaders struggle with AI tools and what needs to change.Follow Steven Webster: Asensei: asensei.ai LinkedIn: linkedin.com/in/stevenwebster Five Principles for AI Leadership: webstersensei.substack.com/p/five-principles-for-ai-leadership AI Leadership and Agentic Office of the CEO: webstersensei.ai Substack: webstersensei.substack.comArtificial Reality: Website: artificialrealitymedia.com Instagram: instagram.com/artificialrealitymedia LinkedIn: linkedin.com/company/artificial-reality-media In-person events: luma.com/artificialreality Host: linkedin.com/in/gfendi

Chris Madden's agency hit 1 billion views in 3 years. He breaks down why AI clipping tools still fail at the most important part of short-form video, how he cloned his voice and now ships content daily without recording himself, and why he believes social media matters more than AI.Chris Madden is the founder of Good Future Media, a podcast editing agency whose clients include All In, This Week in Startups, Pirate Wires, and Solana, and Clipstastic, an AI voice cloning content agency. He spent 10 years teaching English across Thailand, France, Korea, and China before building an unofficial TikTok fan account for the All In podcast that hit 30,000 followers in a month and landed him the role of Head of Content at Twenty VC. He also hosts the Imagine AI podcast with over 150 episodes.Galina Fendikevich asks what's working with AI, what's broken, and what still needs to be built in podcast editing, AI voice cloning, and content production at scale:How ElevenLabs voice cloning lets founders and thought leaders post daily content in their own cloned voice without recording themselves. Chris calls your voice "an extension of your brand" that can now exist across six platforms at once Why AI clipping tools like Opus Clip, Riverside, and Descript still cannot write hooks, add context, or give viewers a reason to stay past the first three seconds. His agency builds every clip by hand in Premiere Pro and CapCut because the human judgment gap has not closed The origin story: how an unofficial TikTok fan account for the All In podcast launched a career at Twenty VC and became a billion-view content agency in three years. Jason Calacanis, the inspiration behind the business, is still hiring human clippers because AI cannot do it either Why Chris wants an "AI brain of your business" that holds full company context in one place. The current friction: you make a decision, change it three minutes later, and your AI still thinks the old goal is active. Memory and context management remain the biggest barriers to using Claude and ChatGPT as real business tools Chris's closing argument that social media distribution is just as important as AI itself. His framework: posting across multiple platforms "increases your surface area for success," and without that distribution layer, no AI tool delivers resultsThis episode is for you if you run a content agency, podcast, or video production company and want a clear breakdown of where AI saves time and where human editors still outperform every tool on the market. It is also for founders, thought leaders, and business owners who want to build a social media presence and post consistently without recording themselves on camera, and for anyone building AI tools for content clipping, voice cloning, lead generation, or recruiting who wants to hear exactly what is failing from someone producing 140 videos a week for the biggest shows in tech.Follow Chris Madden: youtube.com/@ChrisMadden_ instagram.com/chrismadden365 linkedin.com/in/chrismadden365Work with Chris:https://www.cliptastic.ai/ https://goodfuturemedia.com/Artificial Reality: Website: https://www.artificialrealitymedia.com/ Instagram: https://www.instagram.com/artificialrealitymedia LinkedIn: https://www.linkedin.com/company/artificial-reality-media/ In-person events: https://luma.com/artificialreality Host: https://www.linkedin.com/in/gfendi/