Épisodes

  • OpenAI Dev Day 2025: Apps SDK, Agent Kit, Sora 2, ChatGPT as operating system

    OpenAI Dev Day 2025: Apps SDK, Agent Kit, Sora 2, ChatGPT as operating system
    Oct 14 2025
    In this episode of Generation AI, hosts Ardis Kadiu and Petar Djordjevic take you inside OpenAI's third annual Dev Day in San Francisco, breaking down the major announcements that are reshaping how we interact with AI. With ChatGPT now reaching 800 million weekly active users, OpenAI is positioning itself as the operating system of the future. Ardis and Petar, who attended the event in person, discuss three major announcement categories: Apps (native applications running directly in ChatGPT with deep integration), Agent Kit (a visual agent builder with built-in evaluation systems), and new models including GPT-5 Pro, Sora 2 video generation, and cheaper image options. They explore what these changes mean for developers, product builders, and higher education professionals, while sharing their first-hand observations from being in the room with 1,500 developers and AI industry leaders. This episode is essential listening for anyone trying to understand where AI platforms are headed and how to prepare for a future where ChatGPT becomes the hub for all your digital work.Dev Day Experience: San Francisco and the AI Ecosystem (00:00:36)First-time experience attending OpenAI Dev Day in San Francisco with 1,500 attendeesThe unique culture of San Francisco's tech scene and AI billboards everywhereMeeting AI influencers, builders from major companies like Netflix, Facebook, MicrosoftComparing Element451's AI work against world-class builders and feeling competitiveThe optimism and grind culture among new builders and startup foundersThe Three Big Announcement Categories (00:06:32)OpenAI's strategic shift: positioning ChatGPT as an operating systemThree main categories: Apps, Agents, and new ModelsChatGPT reaching 800 million weekly active users (not monthly - weekly)Processing billions of tokens daily across the platformApps in ChatGPT: The Third Try at an App Ecosystem (00:10:05)Native applications running directly in ChatGPT with deep integrationEvolution from plugins (first attempt) to custom GPTs (second attempt) to Apps SDK (third attempt)Launch partners: Canva, Booking.com, Expedia, Figma, Spotify, Khan Academy, Instacart, Uber, TripAdvisorApps can share context with ChatGPT and return custom UI componentsDemo showing Coursera courses, Canva slide creation, and Zillow apartment search all within ChatGPTApps SDK will be available to all developers by end of yearThe Distribution Flywheel and Vendor Lock-in (00:14:53)800 million users creates massive distribution leverage for app makersThe more users work inside ChatGPT, the more context gets centralizedThis strengthens personalization but also increases switching costsChatGPT becoming your memory and general assistantDiscussion of potential for ads and payment systems within ChatGPTUsers becoming more sticky to ChatGPT than to individual app websitesAgent Kit: Visual Agent Builder with Native Evals (00:18:38)Visual agent builder for orchestrating multi-agent workflowsChat Kit for embedding chat interfaces into applicationsNative evaluation system built directly into the platformLive demo: building a full agent for Dev Day conference in 8 minutes on stagePre-built guardrails for PII data and harmful contentConnections to file search, web search, and external systems via MCP protocolSimilar to tools like Zapier, Make.com, and n8n but with embeddable chat widgetsHow OpenAI Uses AI Internally (00:23:44)OpenAI shared three internal use cases at a breakout sessionGo-to-market agent: researches customers before meetings, preps demos, closes the loop after meetingsSupport agent: handles customer inquiries at scale (not outsourced, built in-house)When ChatGPT image generation launched, they got 10 million new users in a dayBuilt-in evals allow systems to improve themselves over time using thumbs up/down feedbackEvals and Prompt Optimization: The Game Changer (00:25:23)Evals explained: non-deterministic outputs require grading systemsEvolution from human graders to LLM gradersOpenAI introducing prompt optimization using the GEPA algorithm (Genetic Pareto)System uses all your data and feedback to automatically improve promptsConnection to DSPY library and the movement toward automated prompt engineeringNot locking users into OpenAI models - can use any model and send traces to the systemComparison with LangSmith and other tracing toolsNew Models: GPT-5 Pro, Sora 2, and Image Mini (00:33:20)GPT-5 Pro now available via API (12x more expensive than standard ChatGPT)Takes minimum 15 minutes per task due to deep reasoning capabilitiesSora 2 and Sora 2 Pro for video generation now in APISora app showing amazing video generation capabilitiesDemo with UK animation studio showing year-long process compressed to minutesGPT Image 1 Mini: 80% cheaper for cost-sensitive, high-frequency tasksEnables personalized images at scale for hundreds of thousands of usersTwo-tier Sora workflow: use smaller model to nail the prompt, then Pro for high fidelityReal-Time Voice Models and Device ...
    Voir plus Voir moins
    56 min
  • Making AI Work: A Buyer’s Guide for Leaders — Part 1

    Making AI Work: A Buyer’s Guide for Leaders — Part 1
    Oct 7 2025

    In this kickoff to a special three-part series titled "Making AI Work: A Buyer’s Guide for Leaders", hosts Ardis Kadiu and Dr. JC Bonilla break down the overwhelming AI landscape in higher education. From general assistants like ChatGPT to workflow-transforming agents, this episode offers a crucial roadmap for institutional leaders wondering what’s hype, what’s helpful, and what’s worth investing in. If you're evaluating ed tech tools or shaping a long-term AI strategy in higher education, this episode is your essential primer.

    00:00 Mapping AI for Enterprise Leaders

    04:10 AI Roadmap: From Foundations to Adoption

    08:29 Understanding AI: Tools vs. Transformations

    12:27 AI Assistants: Fast and Accessible

    15:04 "Enterprise AI Deployments in Institutions"

    18:29 "CRM Copilots and Agents Overview"

    19:43 AI Copilots Revolutionizing EdTech Platforms

    22:55 End-to-End AI Solutions Explained

    26:47 AI-Driven Workflow Automation Tools

    31:49 "Enhancing Systems with AI Workflows"

    36:38 Automated Student Engagement Workflow

    37:46 "AI's Enterprise Value Dilemma"

    43:54 "ROI Correlation with Decision Proximity"

    48:04 Evaluating AI Autonomy Levels

    51:02 "AI Readiness Framework Explained"

    53:16 Evaluating and Comparing AI Vendors

    55:49 Prospects and Students


    - - - -

    Connect With Our Co-Hosts:
    Ardis Kadiu
    https://www.linkedin.com/in/ardis/
    https://twitter.com/ardis

    Dr. JC Bonilla
    https://www.linkedin.com/in/jcbonilla/
    https://twitter.com/jbonillx

    About The Enrollify Podcast Network:
    Generation AI is a part of the Enrollify Podcast Network. If you like this podcast, chances are you’ll like other Enrollify shows too!

    Enrollify is made possible by Element451 — The AI Workforce Platform for Higher Ed. Learn more at element451.com.


    Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
    Voir plus Voir moins
    56 min
  • Breaking Down Vanderbilt’s AI Playbook

    Breaking Down Vanderbilt’s AI Playbook
    Sep 30 2025

    In this special episode of Generation AI, listeners get an exclusive peek into Vanderbilt University’s ambitious and unconventional approach to integrating AI across campus through a recent episode of the AI for U podcast.

    AI for U host Brian Piper speaks with Allen Karns, Chief AI and Technology Officer at Vanderbilt’s Center for Generative AI, who breaks down how Vanderbilt built Amplify—their open-source AI platform—for both administrative and academic use cases. This episode is a masterclass in institutional transformation, innovation culture, and practical AI implementation in higher education.

    Check out AI for U on the Enrollify Network


    - - - -

    Connect With Our Co-Hosts:
    Ardis Kadiu
    https://www.linkedin.com/in/ardis/
    https://twitter.com/ardis

    Dr. JC Bonilla
    https://www.linkedin.com/in/jcbonilla/
    https://twitter.com/jbonillx

    About The Enrollify Podcast Network:
    Generation AI is a part of the Enrollify Podcast Network. If you like this podcast, chances are you’ll like other Enrollify shows too!

    Enrollify is made possible by Element451 — The AI Workforce Platform for Higher Ed. Learn more at element451.com.


    Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
    Voir plus Voir moins
    26 min
  • How People Are Actually Using ChatGPT

    How People Are Actually Using ChatGPT
    Sep 23 2025
    In this episode, Generation AI analyzes groundbreaking research from OpenAI and Anthropic that reveals how AI usage is fundamentally different than expected. Hosts Ardis Kadiu and Dr. JC Bonilla dissect OpenAI's study of 1.5 million ChatGPT conversations, uncovering that 70% of usage is now personal rather than work-related - a complete reversal from initial predictions about enterprise productivity gains. They explore how ChatGPT has reached 700 million weekly active users with 90% of usage now outside the US in less than 3 years (compared to 23 years for the internet), while Claude data shows enterprise users focusing heavily on coding (36% of usage) and autonomous workflows (39% of conversations). The discussion reveals critical implications for higher education: while consumer AI adoption explodes globally with gender parity achieved (52% women users), institutions remain stuck with budget constraints, scattered use cases, and talent retention issues. This episode provides essential insights for education leaders on why the shift toward personal productivity and home-based AI usage creates both untapped opportunities and urgent challenges for institutional AI strategy heading into 2026.OpenAI's Massive ChatGPT Usage Study Overview (00:02:08)Analysis of 1.5 million ChatGPT conversations through NBER working paper700 million weekly active users, most comprehensive AI usage study everCollaboration between OpenAI Economic Research, Harvard economist David Deming, and NBERConsumer plans only - excludes enterprise and API usageSample represents massive scale given ChatGPT's global reachExplosive Growth Patterns and Metrics (00:05:27)Reached 100 million weekly users in under one year (unprecedented speed)Message volume growing even faster than user countAverage user sends 7-8 messages per day (up from 2x in 2024)Cohort analysis shows steady usage for existing users, new users driving intensityGrowth accelerates with each major model releaseGlobal Adoption Outpacing All Previous Technologies (00:08:09)90% of usage now outside North America (achieved in under 3 years)Internet took 23 years to reach same international distributionLower-income countries showing fastest adoption ratesImplications for international marketing and student recruitment strategiesGlobal phenomenon across all economic levelsGender Parity Achievement (00:11:30)Women users increased from 37% (January 2024) to 52% (July 2025)Based on analysis of typically feminine vs masculine namesReflects natural population distribution (50/50 split)Usage patterns now mirror general population demographicsThe Personal vs. Work Usage Revelation (00:13:24)Work-related usage dropped from 47% to only 27%Over 70% of ChatGPT usage is personal/non-work relatedHidden economics of home productivity emerging (not captured in GDP)Similar pattern to mobile device "bring your own device" adoptionEnterprise adoption significantly slower than consumerUsage Intent Categories and Detailed Breakdown (00:16:37)Three main categories: Asking (49%), Doing (40%), Expressing (11%)Practical guidance: 28.8% (top use case)Seeking information: 24.4% (up from 18% year-over-year)Writing: 23.9% (declining as users discover new applications)Multimedia: 7.3% (peaked at 12% after GPT-4o image features)Technical help: ~5%Self-expression: ~5%Specific High-Demand Use Cases (00:19:32)Tutoring/teaching: 10.2% (major opportunity for ed-tech)How-to advice: 8.5% (vertical SaaS potential)Personal writing & editing: 18% (demand for AI co-pilots)Coding in ChatGPT: Only 4.2% (compared to 36% in Claude)Each use case bar represents potential startup opportunity or graveyardClaude/Anthropic Enterprise Usage Analysis (00:27:42)Coding dominates: 36% of Claude usageAutonomous workflows: 39% of conversations (up from 27%)API automation: 77% of business API tasks are full automationMore complex multi-step workflows emergingGeographic usage reflects local economies (NYC: finance, Hawaii: tourism, Massachusetts: science)The Context and Data Bottleneck (00:34:52)Major enterprise bottleneck: Data/context readinessShift from prompt engineering to context orchestration for 2026Context engineering becoming the critical capabilityIntegration with existing platforms determines successOrchestration requires both technology and specialized talentEnterprise AI Economics and Priorities (00:37:26)Companies prioritize capability over cost savingsModel capabilities drive adoption more than pricingBusinesses "lean into automation over cost savings"Not yet highly price sensitive - capacity matters moreBudget lines for AI becoming essential planning itemHigher Education Specific Challenges (00:42:41)Minority of institutions identify as AI leaders75% of CDOs see moderate risk to academic integrityMost exploring scattered use cases vs. campus-wide programsBudget constraints remain primary blockerMarketing and enrollment teams leading adoptionStudent support and advising showing strong use casesTalent retention crisis as AI ...
    Voir plus Voir moins
    48 min
  • From $12 TikToks to $10M Influencers: Oracle Surge, Anthropic Settlement, Google's Veo 3 & Nano Banana Revolution

    From $12 TikToks to $10M Influencers: Oracle Surge, Anthropic Settlement, Google's Veo 3 & Nano Banana Revolution
    Sep 16 2025
    In this episode of Generation AI, hosts Ardis Kadiu and Dr. JC Bonilla explore the accelerating AI infrastructure boom following Oracle's massive 36% stock surge after announcing a $300 billion OpenAI deal. They break down Anthropic's landmark $1.5 billion copyright settlement and what it means for AI training data - companies can train on copyrighted material, but they have to buy it first. The conversation shifts to practical tools as they discuss ChatGPT's new MCP support and Replit Agent 3's autonomous coding capabilities that can work independently for 200 minutes. The hosts then examine the rise of AI influencers making millions through synthetic content, powered by Google's game-changing Nano Banana image editor and Veo 3's new vertical video capabilities that make professional social content creation cost just $12 for a 30-second ad.Opening and UVU Campus Incident (00:00:00)Reflection on September 11th anniversary, 24 years laterDiscussion of tragic incident at University of Utah Valley campusHow institutions manage crisis communicationsAnthropic's $1.5B Copyright Settlement (00:06:32)Court ruling on using pirated materials for AI trainingSettlement details: $3,000 per book for 500,000 illegally obtained booksPrecedent set: AI companies can train on copyrighted material if purchased legallyImplications for other pending lawsuits with OpenAI, Meta, and New York TimesOracle's Stock Surge and Infrastructure Boom (00:10:44)Oracle stock jumps 36% after earnings announcement$300 billion commitment from OpenAI for cloud infrastructureOracle positioning as the "shovels" in the AI gold rushHosting compute for OpenAI, XAI, Meta, and GoogleSignal that AI infrastructure spending is accelerating, not slowingChatGPT Adds Full MCP Tool Support (00:14:51)OpenAI enables MCP (Model Context Protocol) in developer modeMCP as "USB for AI agents" - standardized connection protocolExpanded connectivity beyond limited connector listInternal tools can now expose MCP servers for agent communicationReplit Agent 3: Autonomous Coding Revolution (00:16:37)Third generation agent can work independently for 200 minutesReflective loops for automatic testing and bug fixingCreates multi-step automations similar to Zapier workflowsCan build other agents and complex applications autonomouslySimulates human interaction: clicking, form filling, authenticationThe Rise of AI Influencers and Synthetic Content (00:21:27)Virtual personas making up to $10 million annuallyExamples: Luo Magalo (7.7M followers), Lil Miquella (2M followers)Brands partnering with Samsung, Versace for controlled narrativesAgencies producing synthetic influencers at scaleBalance between fiction/adventure and avoiding deceptionGoogle's Nano Banana: Image Editing Revolution (00:26:12)Transform any image through natural language promptsCharacter consistency for beginning, middle, and end framesUpdate dated content (change "2023" shirt to "2025")Top viral prompts: action figures, different decades, TV showsAvailable in Google Gemini and through APIsVeo 3 Video Generation Goes Social-First (00:28:10)50% price reduction: $0.15 per second for fast generationNew 9x16 vertical format for TikTok, Instagram Reels1080p HD output as standardFull audio integration with voice generation30-second professional ad costs just $12 to producePractical Applications for Higher Ed (00:40:51)Creating personalized content for micro-audiencesUniversity mascots with variations for different demographicsA/B testing at scale for minimal costTransparency and authenticity requirementsBrand ownership and responsibility for synthetic contentTool Recommendations and Alternatives (00:39:12)Midjourney for conceptual images and presentation loopsRunway ML Gen 3 for quick prototypesCling AI 2.1 for lip syncing and motion controlLuma Dream Machine for hyper-realistic animationsLeonardo AI as platform aggregating multiple models including Veo 3 - - - -Connect With Our Co-Hosts:Ardis Kadiuhttps://www.linkedin.com/in/ardis/https://twitter.com/ardisDr. JC Bonillahttps://www.linkedin.com/in/jcbonilla/https://twitter.com/jbonillxAbout The Enrollify Podcast Network:Generation AI is a part of the Enrollify Podcast Network. If you like this podcast, chances are you’ll like other Enrollify shows too! Enrollify is made possible by Element451 — The AI Workforce Platform for Higher Ed. Learn more at element451.com. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
    Voir plus Voir moins
    46 min
  • A16Z's Top 100 AI Apps (5th edition), Google's 4-product surge, vibe coding goes mainstream

    A16Z's Top 100 AI Apps (5th edition), Google's 4-product surge, vibe coding goes mainstream
    Sep 9 2025
    In this episode of Generation AI, hosts JC Bonilla and Ardis Kadiu break down A16Z's fifth annual Top 100 Gen AI Consumer Applications report, revealing how the AI app ecosystem has shifted from experimentation to consolidation. They discuss Google's aggressive entry with four separate products in the rankings, including Gemini's rapid rise to second place behind ChatGPT. The conversation explores how consumer AI has moved beyond novelty to become essential productivity tools, with specialized apps dominating specific use cases like image generation (Midjourney), voice (11 Labs), and the emerging category of agentic coding platforms like Lovable and Replit. The hosts also examine the global dynamics of AI adoption, including the significant presence of Chinese-developed apps and what these consumer trends mean for higher education professionals and their students.Opening and Mediterranean Reflections (00:00:00)Ardis returns from his brother's wedding in TuscanyDiscussion about taking time away from technology and workThe value of slowing down and gaining new perspectives on AI trendsThe A16Z Top 100 Report Overview (00:08:49)Fifth edition of Andreessen Horowitz's annual consumer AI apps reportBased on monthly active users on mobile and web trafficDifference between unique web visitors vs active users explainedFocus on consumer adoption patterns rather than enterprise AIMarket Stabilization and All-Stars (00:11:54)Only 11 new names on web list vs 17 last yearMarket maturity signals with winners consolidating positionsChatGPT reaches 700-800 million weekly active users14 brands consistently dominating across categoriesGoogle's Aggressive Multi-Product Strategy (00:17:00)Gemini takes second place with 12% of ChatGPT's web visitsIntroduction of Nano Banana image editing modelFour Google products separately ranked in top 100Strategic unbundling approach to compete across categoriesThe Rise of Agentic Coding (00:25:47)Evolution from "vibe coding" to "agentic coding"Lovable reaches #22, Replit maintains strong positionIntegration with Supabase for backend developmentReal work being done on these platforms, not just experimentationRegional Dynamics and Chinese Apps (00:29:19)22 out of 50 mobile apps are Chinese-developedDiscussion of China vs rest of world classificationChinese apps being exported globallyAI as a global technology play across regionsYear-Over-Year Changes (00:32:12)Deep Seek's rise and fall (down 40% from peak)Shift from novelty (2024) to utility (2025)Apple's crackdown on ChatGPT copycatsMobile list showing more innovation and newcomersVideo Generation Maturity (00:36:15)Google's VO3 dominates over SoraConsolidation in video generation spaceWorld models like Genie 3 emergingVideo becoming integrated into general assistantsImplications for Higher Education (00:38:01)Students already using multiple consumer AI toolsPattern of utility across companions, creativity, productivityNeed for educators to understand student tool usageRecommendation to explore top apps to understand student behaviorClosing Thoughts (00:41:20)Consumer AI adoption as mirror of society's AI integrationFocus shifting from smartest models to most useful appsEncouragement to test unfamiliar apps from the listPreview of fall conference season and upcoming AI announcements - - - -Connect With Our Co-Hosts:Ardis Kadiuhttps://www.linkedin.com/in/ardis/https://twitter.com/ardisDr. JC Bonillahttps://www.linkedin.com/in/jcbonilla/https://twitter.com/jbonillxAbout The Enrollify Podcast Network:Generation AI is a part of the Enrollify Podcast Network. If you like this podcast, chances are you’ll like other Enrollify shows too! Enrollify is made possible by Element451 — The AI Workforce Platform for Higher Ed. Learn more at element451.com. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
    Voir plus Voir moins
    43 min
  • NotebookLM in Action: From PDF to Podcast in one click, plus mind maps, FAQs & video summaries

    NotebookLM in Action: From PDF to Podcast in one click, plus mind maps, FAQs & video summaries
    Sep 2 2025
    NotebookLM transforms how we work with documents by turning any PDF, YouTube video, or audio file into multiple AI-powered formats with zero prompting required. In this hands-on demo, see how Google's source-grounded AI instantly creates professional 6-minute video summaries, interactive podcasts where you can interrupt and ask questions, visual mind maps for exploring concepts, comprehensive study guides with quizzes, and chronological timelines from your uploaded content. Unlike ChatGPT, NotebookLM never hallucinates because it only references your sources, making it perfect for analyzing research papers, creating course materials, conducting competitive analysis, or synthesizing hours of content into minutes. Real examples show 75% time savings in document analysis, Walter Isaacson using it for biography research, and Element451 creating complete employee onboarding from company documents - proving NotebookLM is the breakthrough tool for anyone drowning in information who needs actionable insights.Back to School Season and Why NotebookLM Matters (00:00:06)Discussion about the start of the school year and its significanceThe class of 2026 as the first to graduate having used ChatGPT throughout collegeIntroduction to NotebookLM as Google's AI-powered research and thinking partnerSetting the stage for practical demonstrations of game-changing featuresWhat Makes NotebookLM Revolutionary (00:05:07)Source-driven design that eliminates AI hallucinationsZero web knowledge - complete focus on your uploaded documentsBuilt on Gemini models but grounded strictly on your sourcesDirect citations and accuracy that ChatGPT can't match for document analysisNotebookLM Interface and Capabilities Overview (00:07:42)Sources panel supporting 50+ documents in pro versionAutomatic summary generation upon any document uploadThe Studio feature that creates multiple asset types instantlySupport for PDFs, audio files, YouTube videos, websites, Google DocsPowerful Real-World Applications at Element451 (00:15:47)Complete employee onboarding packages from company documentsLeadership assessment interpretation from multiple Hogan personality testsCreating interconnected insights from disparate sourcesGenerating training materials that would take weeks in minutesLive Demo: Instant Video Creation from PDF (00:19:30)Loading "Driving Toward a Degree 2025" report demonstration6-minute professional video generated without any promptingMultimedia output with synchronized audio and visual elementsFine-tuning options for specific chapters or topicsRevolutionary Interactive Podcast Feature (00:25:33)AI-generated conversational podcasts from any documentLive interaction where users can interrupt and ask questionsReal-time dialogue with AI hosts about document contentPerfect for learning complex materials through conversationMind Map Navigation for Complex Documents (00:30:29)Visual hierarchy of all document concepts and relationshipsInteractive exploration revealing hidden connectionsDeep-diving capability with automatic source citationsPattern discovery that human readers might missInstant Study Guide and Assessment Generation (00:35:23)Complete study materials created in secondsQuiz generation with answers for immediate useComprehensive glossaries for specialized terminologyFaculty saving entire summers of prep workAdvanced Asset Creation: Timelines and FAQs (00:42:36)Chronological timeline extraction from multiple documentsFAQ generation that took agencies weeks now done instantlyBriefing documents for executive summariesLLM-ready content for chatbot trainingCollaboration and Sharing Capabilities (00:44:38)Share entire notebooks with classes or teamsCreate departmental resources accessible to allPrivacy-protected individual workspacesPerfect for distributed learning environmentsGame-Changing Professional Use Cases (00:46:45)Walter Isaacson analyzing Marie Curie's journals for biographyCreative professionals discovering unconscious patterns in their workFantasy writers maintaining consistency across complex worldsChief Strategy Officer achieving 75% reduction in analysis timeNotebookLM for Learning and Training (00:51:21)Converting 3-4 hours of Kubernetes training into digestible segmentsFinding specific moments in hours of YouTube contentJob interview prep using company websites and descriptionsCompetitive intelligence from competitor materialsWhy NotebookLM Changes Everything (00:56:29)The shift from searching to synthesizing informationElement451's complete adoption for internal processesHow source-focused AI eliminates misinformationThe future of personalized, interactive learning experiences - - - -Connect With Our Co-Hosts:Ardis Kadiuhttps://www.linkedin.com/in/ardis/https://twitter.com/ardisDr. JC Bonillahttps://www.linkedin.com/in/jcbonilla/https://twitter.com/jbonillxAbout The Enrollify Podcast Network:Generation AI is a part of the Enrollify Podcast Network. If you like this podcast, chances are you’ll like other Enrollify shows...
    Voir plus Voir moins
    1 h et 7 min
  • 95% AI pilots fail?, Process is the real AI killer, Buying wins 2x over building, Enter the forward deployed engineer

    95% AI pilots fail?, Process is the real AI killer, Buying wins 2x over building, Enter the forward deployed engineer
    Aug 26 2025
    In this episode of Generation AI, hosts Ardis Kadiu and JC Bonilla examine the widely misinterpreted MIT report claiming "95% of GenAI pilots fail," exploring why this headline misses the real story. While individual employees are finding significant value with AI tools (90% use personal AI regularly), organizations struggle to capture this value at the enterprise level—not because the technology doesn't work, but due to change management, leadership alignment, and implementation challenges. Through Element451's own QBR automation struggles, the hosts illustrate how the gap between impressive demos and measurable business impact stems from organizational readiness, not technological limitations. They discuss why vendor solutions succeed at twice the rate of internal builds (67% vs 33%), introduce Forward Deployed Engineers as the bridge between technology and business context, and explain why back office automation delivers higher ROI than marketing despite budget allocation. This conversation provides practical guidance for higher education leaders on moving from shadow AI productivity gains to true enterprise transformation, emphasizing that the challenge isn't whether AI works—it's how organizations need to evolve to capture its value.AI Deployment Reality Check: The 95% Failure Rate (00:01:35)MIT report reveals 95% of GenAI pilots fail to deliver P&L impactOnly 5% achieve rapid revenue growthDiscussion of how this mirrors Element451's internal experiencesThe difference between pilots, POCs, and actual productsThe Shadow AI Phenomenon (00:02:43)90% of employees using personal AI tools vs enterprise subscriptionsBottom-up adoption through consumer tools like ChatGPTWhy organizations can't measure or control individual productivity gainsThe challenge of enterprise AI adoption vs consumer AIBuilding vs Buying: The Success Rate Gap (00:03:34)Internal build success rate: 33%Vendor purchase success rate: 67%Why vertical solutions outperform generic toolsThe importance of domain expertise in AI deploymentElement's QBR Case Study: When AI Projects Struggle (00:14:18)Quarterly Business Review automation challengesThe gap between data analytics and expert interpretationWhy AI needs embedded best practices and rubricsThe difference between finding patterns and implementing expertiseMarketing vs Back Office: Where Real ROI Lives (00:24:10)Over 50% of AI budgets go to sales and marketingWhy back office automation delivers higher returnsThe binary nature of workflow automation successExamples: fraud detection, application review, transcript analysisThe Forward Deployed Engineer Model (00:35:56)Origin from Palantir's government contractsHow OpenAI uses FDEs for enterprise clientsThe hybrid role: technical expertise + business understandingWhy traditional consultants can't fill this gapThe Unicorn Problem: Finding AI Operations Specialists (00:41:24)Scarcity of people who understand both workflows and AI technologyWhy agencies need to evolve their business modelsThe opportunity for innovative consultanciesElement's challenge in scaling deployment expertiseKey Recommendations for Institution Leaders (00:45:18)Move from bottom-up to top-down AI strategyProperly resource AI initiatives (not just IT side projects)Buy rather than build for 2x success rateLook for vertical solutions with deep domain knowledgeInclude internal champions in deployment projectsFinal Thoughts: Moving Beyond Productivity to Transformation (00:49:31)The shift from individual productivity to enterprise ROIWhy POC success doesn't equal business impactThe importance of AI workflow coverageAccepting that most organizations aren't behind—everyone is struggling - - - -Connect With Our Co-Hosts:Ardis Kadiuhttps://www.linkedin.com/in/ardis/https://twitter.com/ardisDr. JC Bonillahttps://www.linkedin.com/in/jcbonilla/https://twitter.com/jbonillxAbout The Enrollify Podcast Network:Generation AI is a part of the Enrollify Podcast Network. If you like this podcast, chances are you’ll like other Enrollify shows too! Enrollify is made possible by Element451 — The AI Workforce Platform for Higher Ed. Learn more at element451.com. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
    Voir plus Voir moins
    48 min