CXOTalk

Check out Gartner Symposium: https://cxo.news/KGg8XY

===

Tamar Yehoshua is Chief Product and AI Officer at Atlassian, where AI agents are now direct users of the software her team builds. She led product at Google Search and Slack before joining Atlassian, and in this conversation she explains how the company designs products for both people and agents: the teamwork graph that gives an agent the organizational context it arrives without, who stays accountable when a chain of agents goes wrong, how Atlassian measures AI spending with its CFO, and why she never believed agents would kill SaaS.

YOU'LL DISCOVER

✅ Why an AI agent arrives like someone joining your organization with no background whatsoever, and how the teamwork graph (over 200 billion entities) supplies the missing context

✅ The result of Atlassian's A/B tests: 44% higher quality and 48% lower token usage when coding agents use the teamwork graph, versus MCP alone

✅ How Mercedes-Benz built a Rovo agent that cut manual bug triage work for engineers by 85%

✅ The 10x question Atlassian asks before automating a process, instead of settling for a 10 to 20% incremental gain

✅ How Atlassian measures AI value with the CFO: PRs deployed per engineer, features developed, and whether a growing token budget delivers more value

✅ Why vibe-coded replacements for enterprise SaaS "just don't last" at large companies: support, security, and enterprise readiness

✅ The accountability principle for agent chains: the human who builds the first agent stays responsible, with agents increasingly acting as guardrails for other agents

✅ Why Atlassian's AI gateway routes each job to the right model on quality, latency, and cost instead of locking into one model

⏱️ TIMESTAMPS

0:00 Introduction

0:26 Building products for humans and agents

5:50 The teamwork graph and agent context

13:33 Human oversight and agent guardrails

18:26 Moving fast versus getting it right

21:56 Efficiency first, then new capabilities

26:25 Tokens, vibe coding, and SaaS survival

33:00 Agent accountability and the software lifecycle

38:53 Simpler interfaces and visualizing agent work

41:49 How AI reshapes leadership and hierarchy

48:07 Reimagining processes and the Forge ecosystem

50:44 Advice for CIOs and managing AI costs

CXOTalk streams live with audience questions. Join us live and ask the guests your own questions: https://www.cxotalk.com/subscribe

Subscribe to CXOTalk for weekly conversations with leading business and technology leaders.

Get the CXOTalk newsletter: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/atlassian-chief-ai-officer-building-products-for-agents-and-humans

Episode 930 | Recorded August 28, 2026

#CXOTalk #AIAgents #Atlassian #EnterpriseAI #ArtificialIntelligence #AgenticAI #ChiefAIOfficer #ProductManagement #DigitalTransformation #SaaS

Check out Gartner Symposium: https://cxo.news/KGg8XY

========

More than 90% of the code at Snap is now written by AI agents, and a code review agent checks most of it within ten minutes. Saral Jain, SVP, Head of Engineering and CIO at Snap Inc., built the system that makes that safe for nearly a billion monthly users. He explains the golden path of 14 managed agents, the guardrails that account for more than half of the investment, and why accountability still rests with a named human for every change that ships.

YOU'LL DISCOVER

✅ Why Snap funds 14 blessed agents on a golden path instead of letting every team build its own

✅ How Casper turns a Slack conversation into a sandboxed pull request that CodePal reviews before a human sees it

✅ Why more than half of Snap's agent investment goes into evals, code review and rollback rather than into building

✅ The risk-tiering test that separates an internal dashboard from a privacy-sensitive change

✅ Why Saral says the risk is not AI-generated code, it is unowned code

✅ How Snap measures value on outputs, with commits up 75% year over year and severe outages down 57%

✅ Why context beats model 100 out of 100 times, and what context actually means inside a 15-year-old codebase

✅ Four pieces of advice for CIOs, ending with why adoption has to start at the top

⏱️ TIMESTAMPS

0:00 Agents write the code, engineers own it

7:17 Agents as teammates with defined personas

12:52 Systems for agents, features for humans

16:09 How Snap governs thousands of agents

22:39 Invest in guardrails before building agents

29:46 Output metrics and the coordination tax

33:08 Context you can give, judgment you cannot

38:14 The risk is unowned code

44:58 Working with unpredictable AI systems

47:13 Agents beyond engineering, and human skills

50:24 Token spend and model routing

52:35 Advice for leaders and the risks ahead

Subscribe for weekly conversations with leading business and technology leaders.

Get the CXOTalk newsletter: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/snap-agentic-ai-software-development-at-scale

Episode 929 | Recorded August 21, 2026

#CXOTalk #AgenticAI #SoftwareDevelopment #EngineeringLeadership #AICodeReview #Snap #CIO #AIGovernance #DeveloperProductivity #snapchat

Enterprise AI has moved past experimentation and now has to prove its return. Ed Sim, Founder and General Partner of boldstart ventures, ranked the No. 1 seed investor in the Business Insider Seed 100 two years running, sees hundreds of AI startup pitches a year, and writes the first check into companies enterprises buy from years later. He wrote the first check into Snyk and backed Protect AI, which Palo Alto Networks acquired for more than $700 million.

In this conversation, he lays out the three waves of enterprise AI adoption, why rising token costs are pushing companies toward open-weight models and their own hardware, how agent identity and access create a new attack surface, and what separates AI vendors that survive a shakeout from the ones that do not.

YOU'LL DISCOVER

✅ The three waves of enterprise AI: get AI running, get agents running, and the wave happening now, where ROI and tokenomics decide what survives

✅ Why Ed expects dozens of models inside a single enterprise, and the choice he frames as renting intelligence versus owning it

✅ How one portfolio company packaged eight GPUs, CPUs, and a model router into an appliance, routing roughly 10% of queries to the frontier labs and claiming 70% savings per year

✅ Why agents should be granted access at runtime that expires when the task ends, so a breach's blast radius stays contained to one narrow authorization

✅ Cost per outcome as the yardstick: the human doing the task, the AI doing the task, and the human assisted by AI, applied first to discrete work like coding and customer support

✅ A 57-step insurance claims process where the AI was correct 98% of the time and the humans 85%, a gap only visible because every step was recorded

✅ The real difference between open source and open weight models, and why most of Ed's startups now build on open weight models under the hood

✅ Why he argues offense is the new defense, and what the Black Hat sandbox escape means for CISOs planning autonomous defense

⏱️ TIMESTAMPS

0:00 Introduction

0:36 Three waves and the ROI test

3:06 Many models and where startups win

10:32 Who owns access, context, and evaluations

17:21 It's the people, not the architecture

20:04 Measure the outcome, then cut the cost

28:14 Buying talent and changing culture

33:05 Systems of record versus headless agents

36:31 Venture money pivots to robotics and chips

40:11 Open weights and owning your intelligence

44:44 Autonomous attacks need autonomous defense

51:32 Judging vendors and earning enterprise trust

👉 Subscribe for weekly conversations with leading business and technology leaders.

📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com

📝 Show notes, transcript, and summary: https://www.cxotalk.com/episode/top-vc-perspective-where-enterprise-ai-is-headed

Episode 928 | Recorded August 7, 2026

#CXOTalk #EnterpriseAI #AIAgents #AgenticAI #VentureCapital #AISecurity #OpenWeightModels #AIROI #CIO #CISO

Most enterprise AI pilots stall or fail before production, and the blocker is rarely the model.

Nate B. Jones, an AI analyst and advisor who works with Fortune 500 companies and global banks, tells Michael Krigsman that naming an effort a pilot invites small budgets and safe goals.

He explains how to pick a first project that matters to the business, why adoption is roughly 80 percent a people problem, and how to budget AI by cost per completed task rather than cost per token. Recorded live on CXOTalk with questions from the audience throughout.

YOU'LL DISCOVER

✅ Why calling the work a pilot produces smaller budgets, safer goals, and weaker learning

✅ The two starting points Jones gives leaders: get hands-on with the tools yourself, then pick a project where success creates real business leverage

✅ Why adoption follows a bell curve, and what actually moves the middle of the distribution

✅ Why data flow, not the model, is the technical issue that stops initiatives most often

✅ The harness (context, memory, reusable procedures, review gates) treated as company intellectual property

✅ How to compare open weights against frontier models on cost per completed action, including token efficiency between models

✅ Why cost per task keeps falling even as frontier work stays expensive, and how to budget against that

✅ The case for one named owner per agent, and what Jones tells CIOs about shadow AI and cyber defense

⏱️ TIMESTAMPS

0:00 Introduction

0:24 Why pilots fail and where to start

3:30 Adoption is mostly a people problem

8:51 Data, outcomes, and undocumented knowledge

14:19 Learning from pilots and proving value

17:57 The harness and AI fluency

23:59 Why a culture of experimentation wins

28:03 Open weights, costs, and team fluency

33:25 When new model releases matter

38:15 Job fear, AI costs, and accountability

46:41 Agent owners, evals, and production gates

51:03 Advice for CIOs and when to stop

Subscribe for weekly conversations with the business and technology leaders shaping enterprise AI strategy: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/why-ai-pilots-stall-how-to-make-enterprise-ai-work

Episode 927 | Recorded Friday, July 31, 2026

#CXOTalk #EnterpriseAI #AIStrategy #AIAdoption #DigitalTransformation #CIO #AIAgents #Tokenomics #AIGovernance

Are AI agents ready for banking and financial services? Former UBS Group CIO Oliver Bussmann explains the risks to trust and reputation in banking.

This session examines the current adoption landscape of AI agents within the financial sector. With roughly 50% of financial institutions now integrating these tools into their workflows, understanding the operational implications is critical for industry leaders and tech professionals alike.

Bussmann breaks down why financial AI requires a balanced approach. You will learn how firms are navigating the tension between rapid innovation and the need to maintain client trust as they deploy AI agents at scale.

Whether you are managing banking technology or assessing the impact of financial AI, this overview provides context on the real-world challenges facing major institutions today. The discussion highlights the specific reputational hazards that arise when automating sensitive financial processes.

YOU'LL DISCOVER

✅ How copilot use is shifting toward autopilot across back office, IT, and marketing functions

✅ What has to be in place before an agent gets write access to a core ledger: testing, traceability, audit logs, and rollback

✅ Why Bussmann expects audit agents to move into the second and third lines of defense, with PwC and Deloitte already bringing their own

✅ How the risk classification of a use case drives the level of cross-model validation, human verification, and cross-checks

✅ Trust is the asset a bank cannot lose, and Bussmann is waiting for an industry incident driven by hallucination

✅ Why junior software engineer job advertisements are down about 40%, and why you cannot stop hiring juniors you will need as seniors in three to five years

✅ Coding is not the bottleneck; the organizational change required for process redesign is the real constraint

✅ Bussmann is optimistic that agents will run across bank functions within a year, with gains of one, two, or three times in certain use cases against the copilot era's 10 to 30%

⏱️ TIMESTAMPS

0:00 Introduction

0:22 The technology works, the controls lag

6:22 Guardrails first, then measure the gains

9:38 How regulation shapes what agents may do

17:15 Machine learning and high-risk decisions

20:14 Trust is what a bank cannot lose

25:19 Agents are reshaping technology careers

34:26 Risk classification sets the autonomy line

39:29 Customer agents need verified digital identity

45:12 Proving control to regulators and boards

48:23 AI native banks still need people

51:41 Agents in production within a year

Subscribe for weekly conversations with leading business and technology leaders.

Get the CXOTalk newsletter: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/agentic-ai-in-financial-services-former-ubs-and-sap-group-cio

Episode 925 | Recorded July 17, 2026

#CXOTalk #AgenticAI #AIinBanking #FinancialServices #AIGovernance #EnterpriseAI #AIAgents #RiskManagement

Enterprises are running more AI agents than their security teams realize, and attackers only need to be right once. Anand Oswal, EVP of Network Security at Palo Alto Networks, explains how to secure agents across four surfaces: enterprise, SaaS, endpoints, and the browser. With host Michael Krigsman, he covers shadow agent discovery, MCP and browser risks, prompt injection, agent identity, and why a unified platform beats a stack of point products.

YOU’LL DISCOVER

✅ The four agent surfaces every CISO must secure at once: enterprise, SaaS, endpoints, and the browser

✅ Why discovery comes first: you cannot secure agents, models, tools, and plugins you cannot see

✅ The Palo Alto Networks finding that one third of public MCP servers carry takeover level vulnerabilities

✅ How vibe coding agents demand privileged access to local files, terminals, and cloud credentials

✅ How browser agents inherit your session and cookies and can perform identity impersonation

✅ Runtime threats to know: prompt injection, memory poisoning, tool misuse, and model DoS

✅ How MCP and A2A protocols expand the attack surface, and why a centralized AI gateway anchors identity, runtime, and observability controls

✅ The case for zero trust, an AI-driven SOC, and one unified platform over point products, and where Prisma AI fits

⏱️ TIMESTAMPS

0:00 Introduction

0:22 Agent memory poisoning and tool misuse

0:59 Discovering shadow agents across four surfaces

2:32 Vibe coding agents and MCP risk

4:46 Browser agents and session misuse

6:20 Runtime threats and prompt injection

7:17 Agent-to-agent protocols and attack surface

8:04 Agent identity and the control plane

9:16 Centralizing control at the AI gateway

10:23 Zero trust and an AI-driven SOC

11:29 One platform, not point products

Subscribe for weekly conversations with the business and technology leaders shaping the enterprise.

Get the CXOTalk newsletter: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/palo-alto-networks-evp-securing-ai-agents-in-the-enterprise

Episode 924

#CXOTalk #EnterpriseAI #CIO #AIGovernance #AgenticAI #IBM #DigitalTransformation #AIStrategy #AILeadership

CIOs are accountable for AI results but often not in control of how AI is actually used across the business. Andy Baldwin, Senior Vice President of Consulting Offerings and Growth at IBM Consulting, explains how CIOs regain visibility and control as AI moves from small pilots to industrial scale.

He describes the real cost of scaling AI, right-sizing models to cut token cost, governance and observability, cyber and post-quantum risk at the board level, workforce reskilling, and modernizing legacy systems without breaking them.

======

This episode brought to you by Gartner IT Symposium/Xpo™: https://cxo.news/KGg8XY

======

YOU’LL DISCOVER

✅ Why two-thirds of CIOs are accountable for AI but not in full control of how it is used

✅ How IBM runs its own AI program (Client 0) and tracks 60 different models on a single observability layer

✅ Why right-sizing models beats defaulting to an expensive frontier model, the Ferrari-to-the-corner-shop problem that drives token cost

✅ How one AI deployment ran to a $25 million compute cost in six months, then was re-architected down to roughly $2 million

✅ Why AI adoption is a contact sport, not a technology you throw over the fence and hope gets used

✅ Why cyber threats and the post-quantum encryption risk have moved up to the board level

✅ How IBM is reskilling 15,000 to 20,000 people whose skills face declining demand

✅ How to modernize legacy by preserving the system of record while reimagining the engagement layer

⏱️ TIMESTAMPS

0:00 The CIO accountability gap

3:31 Why AI adoption is a contact sport

9:18 Democratization forces a governance rethink

10:31 The real cost of scaling AI

17:01 Writing controls versus enforcing them

19:32 When AI becomes the business model

29:49 From efficiency to reinventing the business

36:07 Quantum and cyber reach the boardroom

41:33 Soft landing or jobs apocalypse

46:59 Proving control and successful pilots

49:54 Modernizing legacy without breaking it

53:06 Accountability and the CIO’s next move

Subscribe for weekly conversations with the business and technology leaders shaping the enterprise.

Get the CXOTalk newsletter: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/ibm-consulting-cios-new-agenda-for-ai

Episode 924

#CXOTalk #EnterpriseAI #CIO #AIGovernance #AgenticAI #IBM #DigitalTransformation #AIStrategy #AILeadership

Can AI startups keep their promise to benefit humanity? Eric Ries explains why business success often leads to corporate "corruption" of the founder's mission.

Eric Ries, creator of the Lean Startup Method, breaks down the inherent tensions between scaling a business and maintaining its core purpose. We examine why so many organizations lose sight of their initial mission as they grow, and what it takes for leadership to stay grounded.

This discussion focuses specifically on how AI company ethics are being tested in the current market. Ries shares his perspective on advising Anthropic, offering a rare look at how a major firm attempts to protect its mission while navigating rapid growth. If you are interested in the intersection of philosophy and corporate strategy, this breakdown offers a practical look at the challenges modern founders face.

Subscribe for weekly business strategy breakdowns, and let me know in the comments: what do you think is the biggest threat to a company's original mission?

YOU'LL DISCOVER

✅ Why corruption means making money without creating value, not breaking the law

✅ The Sol Price story: how FedMart was liquidated, and Costco grew from the same idea

✅ Why shareholder primacy is only about 40 years old, not a law of capitalism

✅ The three-part formula for an incorruptible company: purpose, coherence, integrity

✅ How alternative ownership structures (foundations like Novo Nordisk and Hershey, purpose trusts like Patagonia) make firms far more durable

✅ Eric's idea of financial gravity and why your buying, working, and investing choices matter

✅ The job interview question that can push a company to put its mission in its legal charter

✅ Why Anthropic's public benefit corporation and long-term benefit trust protect its mission

⏱️ TIMESTAMPS

0:00 Why success corrupts good companies

3:14 What corruption really means

5:41 Shareholder primacy is a recent invention

8:22 Sol Price, FedMart, and the founding of Costco

13:08 Who decides which values matter

18:22 Missionaries versus mercenaries

19:44 How Google lost its way

21:38 Governance structures that protect a mission

28:01 Financial gravity and your power

35:25 Red flags when vetting a company

43:00 Why I am optimistic

50:53 Advice for AI founders and Anthropic

Subscribe to CXOTalk for weekly conversations with the business and technology leaders shaping the enterprise.

Get the CXOTalk newsletter: https://newsletter.cxotalk.com

Show notes, transcript, and summary: https://www.cxotalk.com/episode/can-you-build-an-incorruptible-ai-company-a-conversation-with-eric-ries

Episode 923 | Recorded June 26, 2026

#CXOTalk #EricRies #Incorruptible #LeanStartup #CorporateGovernance #ShareholderPrimacy #MissionDriven #Leadership #Anthropic #BusinessEthics

Fewer than 100 companies have scaled enterprise AI from pilots to production to capture great value. Alexander Sukharevsky, who leads QuantumBlack, McKinsey's AI practice, joins Michael Krigsman to lay out the repeatable recipe behind those results and why the winners earn roughly three dollars back for every dollar invested. The conversation covers what capturing AI value really requires, why the CEO and board must own the transformation, and how to lead a hybrid workforce where agents work as colleagues, not tools.

======

This episode brought to you by Gartner IT Symposium/Xpo™: https://cxo.news/KGg8XY

======

YOU'LL DISCOVER

✅ Why fewer than 100 companies captured two-thirds of AI's value, and what they did differently

✅ The repeatable recipe: focus a few domains, ready your data, rewire architecture, and fix the economics

✅ Why AI transformation must be led by the CEO and board, not handed to the CTO or chief digital officer

✅ How to treat AI agents as accountable colleagues, and who stays accountable for the outcomes

✅ Why reinventing a domain beats bolting AI onto an existing process

✅ How the winners pursue cost savings and top-line reinvention at the same time

✅ Why governance and digital trust belong in from day one, with adults in the room on ethics

✅ How expertise and judgment become more valuable as agents speed up the work

⏱️ TIMESTAMPS

0:00 The repeatable recipe for AI value

8:27 Treat agents as colleagues, not tools

13:42 Why the CEO must own the transformation

18:05 From token maxing to value maxing

22:01 Managing a hybrid team of agents

23:30 A flexible architecture for changing models

26:25 Governance and digital trust from day one

30:59 Cost savings versus reinventing the top line

34:46 Human focus, judgment, and accountability

44:12 Redesign workflows instead of bolting on AI

46:24 Careers and apprenticeship in an agent world

50:47 What real CEO ownership looks like

🔔 Subscribe for weekly conversations with the world's top business and technology leaders.

📩 Get the CXOTalk newsletter: https://newsletter.cxotalk.com

💬 Show notes, transcript, and summary: https://www.cxotalk.com/episode/mckinsey-on-agentic-ai-how-to-create-business-value

Episode 922 | Recorded June 19, 2026

#CXOTalk #EnterpriseAI #AI #DigitalTransformation #McKinsey #AIStrategy #AIGovernance #AgenticAI #Leadership

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™: https://cxo.news/KGg8XY

======

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