CXOTalk

Autonomous software development creates a dilemma for leaders in regulated industries: adopt AI coding at scale or fall behind on product velocity without compromising auditability and code quality.

In CXOTalk episode 917, Kris Tokarzewski, Group Chief Technology Information Officer at Vitality, describes how a 14,000-employee multinational insurer is rebuilding its software development life cycle around AI. This episode examines the impact of agentic AI on software development in the enterprise.

Recorded at Blitzy's headquarters, the conversation examines deterministic code generation, Blitzy's infinite code context, context engineering, test-driven development, and the shifting bottlenecks that surface as throughput accelerates.

YOU'LL DISCOVER

✅ Why regulated industries require deterministic, auditable code rather than the probabilistic output most AI coding systems generate
✅ How Blitzy's infinite code context (ingestion of codebases, engineering standards, and business rules) creates high-quality software aligned with compliance requirements
✅ How Vitality reverse-engineers legacy systems with autonomous AI, achieving a measured 5x acceleration over manual methods
✅ Why optimizing end-to-end SDLC throughput matters more than local efficiency at any single stage
✅ How code review of 50,000 to 100,000-line pull requests becomes the next limiting factor, and how AI reviewers close the gap
✅ How test-driven development pairs with autonomous code generation to raise quality and compliance pass rates
✅ How the roles of requirements engineers, software engineers, and product teams converge inside an AI-native SDLC
✅ How to instrument AI spend against velocity, quality, end-to-end throughput, and customer value rather than isolated gains

TIMESTAMPS

0:00 Deterministic code vs. probabilistic AI output
0:14 Meet Kris Tokarzewski, Group CTIO of Vitality
0:32 Why Vitality is modernizing legacy insurance systems
1:30 Event-driven architecture as agentic AI's natural partner
3:00 Building an AI-native software development life cycle with Blitzy
4:28 Throughput optimization versus local efficiency
6:02 Reverse engineering legacy systems and deterministic code generation
9:05 Infinite code context: ingesting codebases, standards, and rules
10:00 Test-driven development with autonomous code generation
10:49 Results: 5x faster legacy reverse engineering
13:17 Product, engineering, and DevOps convergence
15:04 Roles level up: requirements engineers and software engineers
16:18 Reviewing 50,000 to 100,000-line pull requests
17:56 Instrumenting AI spend against business outcomes
19:16 Executive sponsorship for autonomous development
20:16 Advice for CIOs and CTOs adopting AI-driven development

🔔 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-ai-coding-at-global-scale-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 917 | Recorded at Blitzy Headquarters

#CXOTalk #AICoding #AutonomousDevelopment #DeterministicCode #AINativeSDLC #ContextEngineering #InfiniteCodeContext #LegacyModernization #RegulatedIndustries #EnterpriseAI #Blitzy

Agentic AI is reshaping enterprise software faster than most CIOs, CFOs, and vendors are prepared for. Praveen Akkiraju, Managing Director at Insight Partners, joins Michael Krigsman to examine the state of agentic AI in 2026: what works in production, what remains hype, and how sophisticated enterprises are now running more than 1,000 agents at scale. The conversation covers the engineering that separates reliable agents from unreliable ones, the economics of token consumption, and the build-vs-buy calculus facing enterprise buyer4s.

YOU'LL DISCOVER
✅ Why Praveen argues "the agent is actually the harness," and what a harness includes: tools, context, memory, and guardrails
✅ "Jagged intelligence": why state-of-the-art models still fail on basic prompt variations, and the implications for production deployment
✅ How leading enterprises are operating 1,000+ agents and the governance questions that remain unresolved
✅ A bounded vs. unbounded framework for deciding where agent autonomy is realistic and where human approval must stay
✅ Why "token maxing" is consuming annual AI budgets in 90 days, and what CIOs can do about it
✅ How Stampli inserts agentic steps into invoice reconciliation rather than rebuilding the workflow from scratch
✅ Build vs. buy: why front-end workflows favor buying and back-end, data-heavy workflows favor building
✅ The fractional-FTE pricing model emerging for agentic products, and what it means for software economics

⏱️ TIMESTAMPS
0:00 Token maxing and the enterprise AI budget problem
0:23 Model evolution: reasoning, DeepSeek, and the agentic inflection
2:03 What is an agent: models plus harness
4:46 Hype versus reality in agentic AI
8:31 Where agents deliver measurable value today
13:10 Agent negligence, guardrails, and sandboxes
16:06 Data access boundaries: APIs, MCP, and policy files
20:38 Bolt-on agents versus agent-native software
26:53 Human in the loop or autonomous: the operating model question
33:49 Fix your data first, or start now?
41:54 Will agents replace Salesforce and Workday?
47:28 Build vs. buy: front end versus back end
50:45 Token costs and the return of variable-cost software
54:09 Pricing agents as fractional FTEs

🔔 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/agentic-ai-and-the-future-of-enterprise-software-in-2026

🎙️ 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 916 | Recorded April 2026

#CXOTalk #AgenticAI #EnterpriseAI #AIAgents #AIGovernance #CIOStrategy #InsightPartners #EnterpriseSoftware #DigitalTransformation #LLM

Marie Myers, Chief Financial Officer of HPE, explains how she measures business value while deploying agentic AI across a 3,600-person finance organization. Her framework separates direct ROI from indirect value (speed, accuracy, fewer errors) and the operating requirements that make finance AI trustworthy at scale.

YOU'LL DISCOVER
✅ How Myers separates direct ROI from indirect value, including speed, accuracy, and lower error rates
✅ Why determinism was "foundational" for finance AI, and why HPE co-engineered with Nvidia NIMs to achieve consistent answers across half a million data elements
✅ What "human in the loop" means in practice, and why accountability stays with finance leaders
✅ How Alfred (built on Deloitte's Zora platform) moved from transactional workflows to core finance operating rhythms like HPE's weekly ops call
✅ Why clean, reconciled data and a strong data layer are prerequisites for enterprise AI
✅ How HPE redesigned FP&A workflows, centralized the team, and pushed "one source of truth" before layering in agents
✅ How Myers thinks about agile experimentation, stage gates, and when to stop AI investments that will not pay off
✅ Why change management and cultural adoption are often harder than the technology, and how training 3,000+ people was essential

⏱️ TIMESTAMPS
0:00 Measuring AI value beyond hard ROI
3:40 Stage gates, scorecards, and when to stop an AI investment
6:49 "This is a team sport": IT, business, compliance
7:20 Determinism vs probabilism in financial AI
9:38 Alfred, Deloitte Zora, and private cloud (on-premises) architecture
13:04 Human in the loop and limits on agent autonomy
14:31 Highest ROI AI use cases: engineering, marketing, IT
16:23 Where finance sees ROI first: transactional workflows
19:00 "AI slop" and maintaining quality standards
25:32 Data quality and trusted, reconciled financial data
33:49 Redesigning FP&A workflows, "one source of truth"
40:35 Change management is the hardest part of AI

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.
📩 Get the CXOTalk newsletter: newsletter.cxotalk.com
💬 Read show notes and the full transcript: https://www.cxotalk.com/episode/hpes-cfo-making-agentic-ai-work-in-finance

🎙️ ABOUT CXOTALK
CXOTalk features unfiltered conversations with C-suite executives from major companies about AI, digital transformation, and business strategy. Hosted by Michael Krigsman. This is episode 914.

#CXOTalk #HPE #CFO #AIROI #AIinFinance #AgenticAI #AIGovernance #FPandA #FinanceTransformation #EnterpriseAI

Pradeep Mannakkara (CIO) and Ben Mayrides (CISO) of Cvent explain how they govern AI agents at scale across their 5,500-person organization, which now has over 6,000 agents in production. In this fireside chat recorded at a Glean event in NYC, they walk through the AWARE framework developed by Glean's Work AI Institute with Databricks and Palo Alto Networks, and describe the practical tradeoffs of moving fast while managing risk.

The conversation covers agent identity, observability, cultural adoption, CIO/CISO dynamics, and what enterprise-grade AI governance looks like in practice.

You'll discover:
✅ Why traditional IAM and observability controls fail in agentic architectures where agents reason, delegate, and act autonomously
✅ How Cvent deliberately encouraged 6,000 agent creations to build AI fluency before layering in moderation and metrics
✅ The AWARE framework's five pillars: identity, context, guardrails, risk scoring, and ecosystem observability
✅ Why "risk is too high" is never the final answer, only "risk is too high for now"
✅ How Cvent filters AI demand through ROI gates before projects reach security review
✅ Why replacing gut-feel security objections with shared criteria moves the CISO from gatekeeper to business partner
✅ The sandbox-first approach that separates experimentation from production deployment
✅ Why SOC 2 control criteria for AI agents are likely within 18 to 24 months

⏱️ TIMESTAMPS
0:00 Introduction and the AWARE framework
0:34 Core challenges of agent governance
2:43 What agents do for us and to us
4:36 Applying the AWARE framework in practice
7:09 Choosing platforms with built-in controls
9:25 Making governance a cultural shift
11:51 Earning trust through deliberate risk decisions
13:49 Replacing gut reactions with shared criteria
15:20 Managing the CIO/CISO tension
18:54 Shared language for hard tradeoffs
22:01 Go/no-go decisions are never one and done
24:48 Advice for putting AWARE into practice
26:38 Scaling to 6,000 agents

🔔 Subscribe to CXOTalk and hit the bell for new episodes every week.
📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com
💬 Show notes: https://www.cxotalk.com/episode/ai-agent-governance-inside-the-glean-aware-framework-with-cvents-cio-and-ciso

🎙️ 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 913 | Recorded March 10, 2026

#CXOTalk #AIGovernance #AIAgents #CISO #CIO #EnterpriseAI #AgenticAI #AWAREFramework #AICompliance #CyberSecurity

Bill Briggs, CTO of Deloitte, shares findings and advice for Chief Information Officers (CIOs) from the 2026 TechTrends report: 93% of enterprise AI spending goes to technology and tooling, while only 7% of funding goes to culture, change management, and learning. Briggs explains why this imbalance drives failed pilots and runaway costs, and what leaders should do about it.

📌 KEY POINTS
-- Your AI spending ratio is upside down
Enterprises allocate 93% of AI budgets to technology and tooling, while devoting only 7% to culture, change management, and workforce learning. Leaders who invest first in simplifying processes from first principles, before adding AI, consistently produce the strongest returns.

-- Frontline trust in AI sits at 6.7%, and it's costing you
C-suite executives report 70% trust in AI, while entry-level workers register only 6.7%, creating an inverted value chain where the people closest to broken processes stay silent. Organizations can close this gap by declaring intentions upfront and making it safe for workers to experiment openly, rather than hiding behind personal AI tools.

-- Measure outcomes, not agent headcount
Companies broadcasting "tens of thousands of agents" substitute effort metrics for evidence of value; if real business results existed, those numbers would be the headline. Tie every AI initiative to specific operational and financial metrics and kill pilots that result in press releases but no movement that benefits shareholders and employees.

YOU'LL DISCOVER:
✅ Why applying AI to an inefficient process "weaponizes inefficiency" and drives costs through the roof
✅ How trust in AI drops from 70% at the C-suite to 6.7% at the frontline, and why this inverted gap blocks real value
✅ Why hospitals are putting robots on org charts and holding naming competitions for AI coworkers
✅ The specific governance frameworks enterprises need for a workforce of AI agents (modeled on the HR lifecycle)
✅ How inference costs create sticker shock and when to shift from cloud to dedicated hardware
✅ Why Briggs says the CIO's most important skill is now storytelling, not systems architecture
✅ What "success theater" looks like and how to spot it in your own organization
✅ Why 99% of enterprises are fundamentally transforming their IT organizations right now

⏱️ TIMESTAMPS
0:00 Deloitte's CTO: Spend less on technology
0:20 The 93/7 AI spending imbalance
3:59 Why a technologist argues against more tech investment
5:43 State of enterprise AI: 30% reach production scale
8:05 Treating AI deployment like onboarding a coworker
10:29 AI itself means nothing without culture change
13:14 Redesigning work from first principles
16:51 Quantifying AI financial risk and token economics
20:03 Inference costs, shadow IT, and runaway bills
23:14 The trust gap: 70% at the top, 6.7% at the bottom
26:47 Governing a workforce of AI agents
32:15 Success theater vs. real business metrics
37:37 Responsible deployment, guardrails, and OpenClaw lessons
42:37 How AI is transforming the CIO role
46:05 Why storytelling is the CIO's most important skill
50:02 Human times machine: the essential equation

🔔 SUBSCRIBE for weekly conversations with global technology and business leaders who speak candidly about the strategies behind AI, transformation, and organizational change.
📩 Get notified about upcoming episodes and exclusive insights: https://newsletter.cxotalk.com
💬 Read show notes and get the transcript: https://www.cxotalk.com/episode/deloitte-cto-on-the-ai-investment-trap-cio-advisory-2026

🎙️ 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 912 | Recorded March 15, 2026

#CXOTalk #AIStrategy #EnterpriseAI #DigitalTransformation #Deloitte #CIO #AIGovernance #TechTrends2026 #AIInvestment #AgenticAI