Mayfield CXO Insights Call:
Rebuilding IT for the Agentic Enterprise
with Taj Singh, CoreWeave
Tuesday, November 3rd,
10-11am PT
Hosted by Mayfield
$3B in AI
We’re the first institutional investor �for early-stage AI founders. �70% of our investments begin at inception.
INVESTMENT
TRACK RECORD
120
Our track record reflects �a simple belief: great companies �are built by people first.
IPOs
225+
M&As
2
Proprietary and Confidential. Do not distribute.
Semis
Agentic Teammates
Applications
Middleware
Data
Models
Building the AI First Enterprise
Where Mayfield Is Investing
Proprietary and confidential
4
Enterprise AI 2030
The Mayfield Annual CXO Survey
I’d personally value your input in this year’s Mayfield CXO Survey.
The Mayfield CXO Survey, “Enterprise AI 2030” explores where organizations stand today and what leaders expect by 2030.
Please Take the 10-minute survey
Individual responses will remain confidential; findings will be reported in aggregate.
As a thank-you, Mayfield will make a $500 charitable donation on your behalf to a cause you select: university research and innovation, college scholarships, childhood hunger relief, initiatives led by and for women, or cancer research.
Scan to take the 10-minute survey
Individual responses remain confidential. Findings will be reported in aggregate.
Proprietary and confidential
5
Rebuilding IT for the Agentic Enterprise:
Taj is a product-centric CIO leader at CoreWeave, building a model AI-native IT team that delivers productivity and business multiplier effects. Before moving into enterprise IT leadership, Taj built and shipped enterprise and AI products at Uber, Microsoft/Xbox, and [24]7.ai. That product background shapes an approach to IT as a business outcome multiplier, not a cost center. Taj mentors emerging technology leaders. He serves on the advisory boards of Asana and Atlassian and invests in and advises AI startups on strategy and infrastructure.
Taj Singh
Head of IT Strategy &
Transformation
Imagination asymmetry
1 / 8 - AI transformation
When everyone has the same information, advantage comes from what you imagine
We're moving from information (and execution) asymmetry to imagination asymmetry.
Didier
Creation is not the bastion of software developers any more
Five steps to AI transformation
2 / 10 - AI transformation
Connected steps that evolve together, not a waterfall
1
Understand your starting point
2
Set goals for literacy, fluency, and ROI
3
Choose your technology strategy
4
Establish the operating model and governance
5
Create transformational value
Your starting point
3 / 10 - AI transformation
Company context shapes the pace, investment, and approach
AI native
Freedom to choose tools, people, and processes
Established enterprise
Legacy commitments and change at scale
Small enterprise
Focused priorities and limited capacity
Hybrid enterprise
Different businesses moving at different speeds
Profiles overlap. Founder-run companies are like AI natives.
KEY DRIVERS
Leadership commitment
Data fluency
Adaptability
AI transformation goals
4 / 10 - AI transformation
Targets for literacy, fluency, and ROI give the company a direction
Literacy
Broad participation
MEASURES
Prompts per user
Employee participation
Fluency
Creation within teams
MEASURES
Agents, skills, and plugins created by users
ROI
Useful agents, repeated use
MEASURES
Repeat use in workflows
Measured business outcomes
Measure # of transformational agents
Set a target, owner, and timeframe for each stage
AI transformation goals dashboard
5 / 10 - AI transformation
An example layout for tracking adoption and spend.
MONTHLY ACTIVE USERS
Unique users across all tools
AI PENETRATION
Active users ÷ headcount
INTERACTIONS / USER
Depth of everyday use
SPEND / USER
Total AI spend ÷ active users
ADOPTION
Monthly penetration across tools
Penetration by department
Trend for each tool
SPEND
Spend by provider over time
Spend by department
Token usage
Technology strategy
6 / 10 - AI transformation
Four paths to capability, supported by an enterprise platform stack
FOUR PRONG TECH STACK
1
Embedded AI (Compressed SoRs)
Workday
Oracle
Salesforce
2
Functional AI applications
Harvey
3
Enterprise platforms and AI labs
Glean
Workato
Claude
ChatGPT
Perplexity
4
Custom AI development
Built in-house with Python and frameworks such as LangGraph and LangChain
PLATFORM STACK
Agent creation platform
Agent orchestration
Agent hosting
Agent registry
A2A (agent-to-agent)
Agent testing and evaluation
Enterprise context
AI gateway
Connectors
API key management
AI security
Models
Data platform
Analytics and monitoring
…
Ema
Technology strategy
6 / 10 - AI transformation
Four paths to capability, supported by an enterprise platform stack
FOUR PRONG TECH STACK
1
Embedded AI (Compressed SoRs)
Workday
Oracle
Salesforce
2
Functional AI applications
Harvey
3
Enterprise platforms and AI labs
Glean
Workato
Claude
ChatGPT
Perplexity
4
Custom AI development
Built in-house with Python and frameworks such as LangGraph and LangChain
PLATFORM STACK
Agent creation platform
Agent orchestration
Agent hosting
Agent registry
A2A (agent-to-agent)
Agent testing and evaluation
Enterprise context
AI gateway
Connectors
API key management
AI security
Models
Data platform
Analytics and monitoring
…
Ema
Established
Manual / WIP
Operating model
7 / 10 - AI transformation
Shared enterprise capability, with accountability in every business
Central hub
Accountable: CIO
AI transformation owner
FDEs (IT)
REMIT
Overall enterprise AI
Business spoke
Repeats in each business unit or function
Accountable: CXO
AI owner
Agent Operations Engineers
SPOKE RESPONSIBILITIES
Registry
Risk assessment
Graduation to hub
Evaluations
Strategic partnership with the People team is key
Enablement / Hackathons + Standards + Tech + Scaled Agents
Governance
8 / 10 - AI transformation
Clear decision rights, with a continuous feedback loop
Continuous
feedback loop
AI governance council
Policy, priorities, risk, and economics
Set direction
AI technology council
Architecture, integrations, and model decisions
Guide execution
AI ambassadors
Grassroots learning and feedback
Share what works
Security, privacy, and compliance throughout
Few agents are transformational
9 / 10 - AI transformation
Our estimate: 10% or fewer. Most agents are automations like RPA and chatbots
ILLUSTRATIVE: 1 IN 10
Automations
RPA-style tasks, chatbots, and assistants. Useful, but the workflow stays the same.
Transformational agents
Run a whole workflow across systems, and change how the work gets done.
EXAMPLES
Hiring integration agent
Warehouse receiving agent
Transformational value
10 / 10 - AI transformation
Product and design thinking guide process redesign and ROI
Product thinking
The problem worth solving
Design thinking
The experience people need
Outcome-based Process
A better way to do the work
Transformational agents
The multiplier comes from rethinking the whole workflow