CXOTalk

Agentic AI has taken off in software engineering, but most CIOs still cannot make agents work in everyday knowledge work in the enterprise. Aaron Levie, co-founder and CEO of Box, explains why that gap exists and what enterprises must change to close it.
Drawing on what Box sees across its enterprise customer base, including 68% of the Fortune 500, Levie covers data access, verification, budgets, architecture, and the new roles required to realize real value from enterprise AI agents.
======
This episode is brought to you by Gartner IT Symposium/Xpo™.
Ready to scale agentic AI from pilot to production? Join top CIOs and IT executives at Gartner IT Symposium/Xpo, taking place this October 19th through the 22nd in Orlando, Florida. Over 300 Gartner analyst-led sessions will cover top priorities shaping IT—from AI value, governance, and cybersecurity to cost optimization, IT operating models, and beyond. Get practical, actionable insights—and connect with peers tackling the same challenges you are.
Secure your spot today at gartner.com/us/symposium.
======
YOU'LL DISCOVER
✅ Why agentic coding raced ahead while knowledge work agents lag, across three properties: text based work, verifiability, and data access
✅ The "AI psychosis" pattern Levie says makes CEOs overestimate agents, and why distance from the last mile of work distorts executive judgment
✅ Why you should retry a failed AI project roughly every six months as frontier models keep improving
✅ The forward-deployed engineer role, internal and external, and why it becomes essential to enterprise AI adoption
✅ Why your IT and data architecture, not the model you pick, often determines what you actually get from agents
✅ The end of venture-subsidized tokens, and why the line of business, not just IT, now has to own the AI budget
✅ Why Levie says you should not vibe-code core systems of record like ERP or CRM, and where agent value actually accrues
✅ Value maxing versus token maxing: how to judge AI ROI and avoid a surprise overnight token bill

⏱️ TIMESTAMPS
0:00 The promise of agentic coding
5:11 Why knowledge work resists agents
8:52 The AI psychosis trap for CEOs
14:57 Be ambitious, then retry in six months
17:25 The rise of the forward-deployed engineer
21:09 Frontier models need your data architecture
27:14 The end of subsidized tokens
31:18 How knowledge workers should prepare
36:37 Where software value shifts
39:03 Reimagining workflows around abundance
43:03 Value maxing versus token maxing
49:46 Advice for CIOs

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.
📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com
💬 Read the show notes and episode summary: https://www.cxotalk.com/episode/box-ceo-aaron-levie-cio-advice-on-agentic-ai-and-the-enterprise

#CXOTalk #AaronLevie #Box #EnterpriseAI #AIAgents #AgenticAI #DigitalTransformation #CIO #KnowledgeWork #AIStrategy

Most enterprises are renters, not owners, of their technology and AI. Raffi Krikorian, Chief Technology Officer of Mozilla, explains why dependence on a handful of closed model providers means losing control over model behavior, pricing, and your own data.

In CXOTalk episode 920, Krikorian lays out where open-source AI actually wins in the enterprise, how lock-in happens quietly, and what CIOs and CTOs should do about it now. Krikorian draws on his experience building infrastructure at Twitter and running the self-driving division at Uber to ground the discussion in real engineering and economic tradeoffs, not hype.

YOU'LL DISCOVER
✅ Why 85% of enterprises believed they could switch AI vendors, but only about 30% actually could when they tried
✅ The "renters vs. owners" framing and what it means to control your AI destiny
✅ Why Krikorian wants data "protected by architecture, not legal handshakes"
✅ How Pinterest reportedly saved on the order of $10 million in a single quarter by switching from closed to open models
✅ Why IT is becoming "the HR team for agents," and the read/write "dangerous triangle" of agentic permissions
✅ The case for recording your prompts and running your own evaluations instead of trusting public benchmarks
✅ Why roughly 70% of enterprise GPUs sit idle, and the missing "LAMP stack for AI" that could put them to work
✅ How closed "validation machines" can quietly steer answers toward sponsored outcomes

⏱️ TIMESTAMPS
0:00 Renters vs. owners: who controls enterprise AI
2:26 The risks of depending on closed model makers
6:23 How lock-in happens and where open source fits
9:53 Regression testing and building your own evals
13:24 Pricing instability and the post-IPO cost question
23:31 Governance: IT as HR for AI agents
32:38 Can a small organization own its AI stack end-to-end?
38:47 Validation machines, trust, and sponsored answers
43:39 Keeping humans at the center, not in the loop
47:23 Can open source beat big tech in AI?
51:39 Inside Mozilla.ai: Otari, CQ, Octanus, Thunderbolt
55:21 The "rebel alliance" strategy

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.
📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com
💬 Read the show notes, summary, and transcript: https://www.cxotalk.com/episode/mozilla-cto-open-source-ai-agents-and-the-fight-for-control

🎙️ ABOUT CXOTALK
CXOTalk features unfiltered conversations with C-suite executives from major companies about AI, digital transformation, and business strategy. Hosted by Michael Krigsman.
Episode 920

#CXOTalk #EnterpriseAI #OpenSource #AIGovernance #CIO #Mozilla #DigitalTransformation #AIStrategy #VendorLockIn #AgenticAI

AI agents are entering enterprise AI faster than CIOs can govern them. Line-of-business users are vibe-coding their own tools, agents are operating with employee credentials, and foundation models are changing under running systems.
In CXOTalk episode 919, Anthony Scriffignano, PhD, a prominent data scientist, and Tim Crawford, a strategic advisor to CIOs at the world's largest companies, examine what enterprise AI governance, shadow AI, and agentic risk require of technology leaders today. The discussion grounds the AI agent conversation in practical decisions: what to keep from established IT governance, what is genuinely new, and where the CIO role must evolve.
YOU'LL LEARN:
✅ Why traditional regression testing breaks when foundation models, training data, and environments all change at once
✅ How shadow AI and vibe-coding by non-developers expand the threat paradigm beyond the enterprise perimeter
✅ Why HR-style policies do not transfer to AI agents, and what changes when super-agents call sub-agents through an orchestration layer
✅ Specific controls for shadow AI: sandboxes, token counting, personal Identifying Information (PII) guardrails, and watching for value leaving the organization
✅ Red, blue, and green teaming for autonomous agents, including why red teams need a defined target list, not a license to break things
✅ The three governance layers CIOs must now reconcile: user role-based access controls (RBAC), agent governance, and knowledge governance, across ServiceNow, Salesforce, and SAP
✅ When human in the loop is meaningful and when it becomes theater, including the limits of audited-sample review at machine speed
✅ How the transformational CIO mindset differs from the traditional one, and why business depth is now the prerequisite skill
⏱️ TIMESTAMPS
0:00 AI agents are running wild: framing the problem
3:11 From automation to autonomy: how CIOs should reframe risk
5:21 What old governance disciplines still apply, and what is new
6:12 Shadow AI, vibe coding, and the limits of control
9:11 Practical controls: sandboxes, token counting, PII guardrails
11:53 Why HR policies do not work for AI agents
15:24 Regression testing for misuse and misadventure
18:43 The aspiring CIO: traditional vs. transformational mindset
21:07 Disciplined red, blue, and green teaming
23:30 When mandatory automation becomes the only option
32:03 Human in the loop: meaningful or theater?
34:09 What AI governance actually looks like in practice
38:10 New roles: context engineers, AI FinOps, and value frameworks
40:30 Talent and jobs inside IT: what changes

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.
📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com
💬 Read the show notes: https://www.cxotalk.com/episode/cio-playbook-agentic-ai-in-the-enterprise
🎙️ ABOUT CXOTALK
CXOTalk features unfiltered conversations with C-suite executives from major companies about AI, digital transformation, and business strategy. Hosted by Michael Krigsman.
Episode 919
#cxotalk #ShadowAI #AIAgents #AIGovernance #AgenticAI #CIO #EnterpriseAI #DigitalTransformation #AIRisk #CIOLeadership

Enrique Ibarra, CIO and Head of Business Transformation at GNP, Mexico's largest insurance company, walks through an enterprise-scale pilot of autonomous software development involving roughly 1,000 internal and external developers. The episode examines how agentic AI changes developers' roles from creators to editors and orchestrators.

In CXOTalk episode 918, Ibarra explains why AI co-pilots alone were insufficient to modernize a 20-year-old mainframe system, how GNP evaluated the Blitzy autonomous development platform across four real-world use cases, and how developer roles are shifting from creators to editors and orchestrators. The episode covers legacy modernization, enterprise AI adoption, change management, measurable results, and the two-year roadmap to retool the full engineering organization.

YOU'LL DISCOVER

✅ The CIO's phased human-in-the-loop playbook: target high-effort, low-risk friction points first (documentation, test suites, version upgrades)
✅ Measured outcomes: 5 to 10X engineering velocity, near-100% autonomous completion on language upgrades, roughly 80% on frontend modernization
✅ Why GNP's 20-year-old mainframe system forced a modernization decision tied to cost and the coming COBOL talent shortage
✅ How the pilot was structured across four use cases: Java 8 to Java 21 migration, Angular frontend upgrade, new feature build, and security vulnerability remediation
✅ Why autonomous platforms differ from co-pilots, and when to use each (Blitzy for heavy lifting, IDE-based co-pilots for the final 20%)
✅ How to encode technical, security, and architectural guidelines as prompt inputs rather than post-hoc review
✅ The change management approach that converted skeptical developers into active users within weeks
✅ Strategic payoff: shipping new insurance products in weeks rather than months, and shifting IT from maintaining the business to dictating market pace

TIMESTAMPS

0:00 Introduction and headline results
0:39 Why GNP needed to modernize a 20-year-old mainframe system
1:15 From coding co-pilots to an autonomous platform
2:36 Designing the four-use-case pilot
4:26 Autonomous platforms versus vibe coding
5:49 What autonomous development means in practice
7:24 Encoding security and governance as prompt inputs
8:24 Results: velocity, autonomy rates, and the final 20%
10:16 How developer roles and daily work change
11:19 Managing developer skepticism and change resistance
12:25 Advice for CIOs: the phased human-in-the-loop playbook
13:34 Strategic business benefits and first-to-market product launches
14:58 Rolling out across seven teams and a two-year horizon
16:34 Final advice for engineering leaders getting started

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.
📩Get the CXOTalk newsletter: https://newsletter.cxotalk.com
💬 Read the show notes: https://www.cxotalk.com/episode/autonomous-software-development-at-enterprise-scale-inside-a-1-000-developer-pilot-with-blitzy

🎙️ ABOUT CXOTALK
CXOTalk features unfiltered conversations with C-suite executives from major companies about AI, digital transformation, and business strategy. Hosted by Michael Krigsman. Episode 918

#CXOTalk #AutonomousSoftwareDevelopment #Blitzy #AgenticAI #EnterpriseAI #CIO #AICodeGeneration #LegacyModernization #DigitalTransformation #SoftwareEngineering

AI swarms are now considered the most dangerous influence weapons ever created, actively fabricating grassroots consensus and corrupting enterprise AI training data through disinformation.
Daniel Thilo Schroeder, Research Scientist at SINTEF, and Jonas R. Kunst, Professor at BI Norwegian Business School, co-authored a study with 22 authors published in Science that maps this threat. They explain how AI swarms operate without human oversight, why traditional detection methods fail, and what governments, platforms, and business leaders must do to fight back. This is CXOTalk episode 915.

YOU'LL DISCOVER
✅ How AI swarms shift from central command to emergent hive behavior with decreasing human oversight
✅ Why AI-generated social media messages now pass the Turing test, rendering individual message detection obsolete
✅ The persona-centric architecture: how single AI agents coordinate behavior across email, X, Bluesky, and Facebook simultaneously
✅ How swarms fabricate synthetic consensus by hijacking human conformist psychology
✅ The perverse incentives of social media business models that profit from AI swarm engagement metrics
✅ How AI swarms poison LLM training data, causing future models to output manipulated facts as objective reality
✅ The proposed Distributed AI Influence Observatory for decentralized threat intelligence sharing
✅ Why malicious actors can deploy self-optimizing AI swarms from a bedroom using existing multi-agent frameworks

⏱️ TIMESTAMPS
0:00 The Shift from Bot Networks to AI Swarms
2:00 Why Cheap AI Inference Enables Long-Term Influence Campaigns
4:30 Autonomous Coordination and Emergent Hive Behavior
7:00 Persona-Centric Agents Across Multiple Platforms
8:30 Weaponizing Disinformation to Fabricate Synthetic Consensus
14:15 How AI Swarms Corrupt LLM Training Data
18:00 Why Individual Message Detection No Longer Works
23:00 The Research Frontier: Coordination Pattern Detection
27:00 Platform Business Models and Perverse Incentives
32:00 Building Defenses: The AI Influence Observatory
39:00 Corporate Risks: Fabricated Boycotts and Targeted Harassment\
46:00 Can It Be Stopped? The Arms Race Democracies Must Join

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.
📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com
💬 Read the show notes, summary, and transcript: https://www.cxotalk.com/episode/how-ai-swarms-weaponize-disinformation

🎙️ ABOUT CXOTALK
CXOTalk features unfiltered conversations with C-suite executives from major companies about AI, digital transformation, and business strategy. Hosted by Michael Krigsman.

#CXOTalk #AISwarms #Disinformation #InformationWarfare #Cybersecurity #AgenticAI #TechPolicy #EnterpriseRisk #Democracy #InfluenceOperations