Jesse Zhang is a Chinese-American technology entrepreneur and co-founder and CEO of Decagon, one of the fastest-growing AI agent companies in the enterprise software market.

Decagon is building AI agents that act as an "AI concierge" for businesses, handling customer-service interactions across chat, email, and voice while connecting directly to the systems businesses use to resolve customer issues.

Zhang is particularly notable because he reached this level of success at a remarkably young age. Before Decagon, he founded Lowkey, a gaming-focused startup that was acquired by Niantic in 2021. He then turned his attention to enterprise AI, co-founding Decagon in 2023.

By early 2026, Decagon had raised another $250 million, taking its valuation to approximately $4.5 billion. The company subsequently reported $100 million in annualized revenue in July 2026, according to Sacra's estimates.

For CEO-Watch, Zhang represents an increasingly important generation of founders who have moved quickly from consumer startups into AI-native enterprise software.

A Young Founder With an Unusually Technical Background

Zhang's interest in technology began well before Decagon.

He studied Computer Science at Harvard University, graduating in just three years. Before and during university, he participated in highly competitive mathematics and computer-science programs, including the USA Mathematical Olympiad Program and other national STEM competitions.

His technical background includes:

Computer science
Algorithms
Artificial intelligence
Mathematics
Software engineering
Data analysis

Before founding companies, Zhang interned at organizations including Google, Citadel, Hudson River Trading, Intel, Kensho, and EDO.

This background gave him exposure to both software engineering and highly quantitative environments.

Lowkey: Zhang's First Startup

Zhang's first major startup was Lowkey, a gaming and social platform.

He founded Lowkey in 2018 and served as its CEO.

The company was backed by Y Combinator and Andreessen Horowitz, giving Zhang early exposure to Silicon Valley's venture ecosystem.

Lowkey focused on helping gamers create and share content around their gaming experiences.

The company's product sat at the intersection of:

Gaming + Social + Content Creation

This was an important learning experience for Zhang because it forced him to think about:

Consumer behavior
Product engagement
Community
Growth
User experience
Viral distribution

The Niantic Acquisition

In December 2021, Niantic acquired Lowkey.

Zhang subsequently joined Niantic and worked on its social team.

The acquisition gave Zhang experience operating inside a larger technology company after spending several years as a founder.

That combination—startup founder + scaled technology company experience—would later influence how he approached Decagon.

From Consumer Social to Enterprise AI

After Lowkey, Zhang made a major strategic shift.

Instead of building another consumer application, he moved toward enterprise artificial intelligence.

He co-founded Decagon with Ashwin Sreenivas in 2023.

Sreenivas became Decagon's co-founder and CTO, while Zhang took responsibility as CEO.

The founders believed that recent advances in large language models had finally made it possible to build software that could perform complex customer-service work.

The Problem With Traditional Customer Service

Customer service has historically depended on a large human workforce supported by software.

A typical interaction looks like:

Customer → Support agent → CRM / knowledge base / internal tools → Resolution

The problem is that the human agent has to navigate multiple systems to resolve even relatively simple requests.

Customers may need help with:

Returns
Cancellations
Account changes
Lost cards
Reservations
Billing
Product questions
Shipping
Refunds

Decagon's thesis is that AI agents can increasingly handle these processes themselves.

The "AI Concierge"

Decagon describes its product as an AI concierge.

The distinction is important.

A traditional chatbot primarily answers questions.

A Decagon agent is designed to:

Understand → Reason → Use tools → Take action → Resolve the customer's issue

For example, instead of telling a customer how to return an item, an agent could potentially:

Identify the order
Verify the customer
Check eligibility
Initiate the return
Generate the required information
Confirm the next steps

The goal is therefore not simply automated conversation.

It is automated resolution.

The Agentic Customer-Service Market

Decagon emerged into a highly competitive AI customer-service market.

Major competitors and adjacent companies include:

Sierra
Salesforce
Intercom
Zendesk
Ada
Gorgias
PolyAI
Various AI-agent startups

But Zhang has positioned Decagon around a more ambitious proposition:

AI agents should become the primary interface between brands and their customers.

Decagon's March 2026 internal memo states this vision explicitly: every major brand could eventually have an AI agent serving as its primary customer-facing interface.

The Importance of AI Agents Taking Action

Zhang's strategy is based on a fundamental distinction between generative AI and agentic AI.

Generative AI:

"Here's the answer."

Agentic AI:

"I'll take care of it."

This is a much larger market opportunity.

If an AI can actually execute workflows, it can potentially replace substantial amounts of repetitive operational work.

That makes the economic value of the software much greater than simply reducing the time employees spend reading and writing messages.

Decagon's Agent Operating Procedures

One of Decagon's key product concepts is Agent Operating Procedures, or AOPs.

AOPs allow businesses to define how their AI agents should behave using natural language while technical teams maintain control over the underlying systems and code.

The concept is similar to combining:

Business rules + AI reasoning + software execution

This gives customer-experience teams greater control over the behavior of agents without requiring them to become AI engineers.

Decagon says this approach has become one of the differentiators of its platform.

Putting the Customer Experience Team in Control

This reflects an important product philosophy.

Traditional enterprise AI often requires technical teams to configure everything.

Decagon attempts to move some of that control closer to the people who actually understand customer interactions.

A CX team can describe:

What the agent should do
What it shouldn't do
When it should escalate
How it should communicate
What policies it should follow

Engineering teams maintain the underlying infrastructure and integrations.

This creates a bridge between:

Business expertise

and

AI engineering.

Rapid Enterprise Adoption

Decagon's growth has been unusually fast.

In June 2025, the company announced that its AI agents were already being used by companies including:

Hertz
ŌURA
Duolingo
Bilt
ClassPass
Notion

The platform was supporting interactions across chat, email, and voice.

By the end of 2025, Decagon had added more than 100 global enterprise customers, including Avis Budget Group, Block, and Deutsche Telekom.

$131M Series C

In June 2025, Decagon raised a $131 million Series C.

The round was co-led by:

Andreessen Horowitz
Accel

The financing valued Decagon at approximately $1.5 billion.

It brought the company's total funding to approximately $231 million at the time.

The milestone was particularly remarkable because Decagon had only emerged from stealth roughly a year earlier.

The $250M Series D

Decagon's growth accelerated further.

In January 2026, the company announced a $250 million financing round, led by Coatue Management and Index Ventures.

The financing tripled Decagon's valuation to approximately:

$4.5 billion

in less than six months.

The company said the round followed a year in which more than 100 new global enterprise customers joined the platform.

For a company founded only in 2023, the speed was extraordinary.

$4.5 Billion Valuation

The $4.5 billion valuation placed Zhang among the most valuable young founders in enterprise AI.

Forbes reported in February 2026 that Decagon had raised more than $500 million in total funding and reached the $4.5 billion valuation.

Decagon subsequently conducted an employee tender offer at the same $4.5 billion valuation, giving employees an opportunity to gain liquidity from their equity.

That was an unusual move for a company at such an early stage.

$100M Revenue Milestone

By July 2026, Sacra estimated that Decagon had reached approximately $100 million in annualized revenue.

The same analysis estimated that revenue had grown from roughly $44 million at the end of 2025.

According to Sacra, Decagon added more than 100 global enterprise customers during 2025 and continued to expand rapidly in 2026.

While this is an estimate rather than a company-reported financial figure, it illustrates the scale of Decagon's commercial growth.

From Chat to Voice

Decagon has expanded beyond text-based customer support.

Its agents operate across:

Chat + Email + Voice

This is strategically important.

A customer shouldn't have to care which communication channel they're using.

The underlying AI should understand the customer regardless of whether the interaction happens through:

Website chat
Email
Phone
Messaging

That supports Zhang's broader vision of an AI concierge rather than a simple chatbot.

Proactive AI Agents

In March 2026, Decagon introduced Proactive Agents, including capabilities around user memory and outbound voice.

This changes the relationship between a business and its AI agent.

Traditional support:

Customer has a problem → Customer contacts company

Proactive support:

AI detects an opportunity → AI contacts customer → AI helps resolve the issue

That creates a potentially much larger role for AI in customer experience.

The agent isn't simply responding.

It can proactively engage.

The Agent as the Interface

Zhang's longer-term vision is even broader.

Instead of customers navigating a company's website, app, or help center to find the correct workflow, they may simply interact with the company's AI agent.

For example:

"I need to change my flight."

Instead of navigating:

Account → Reservations → Flight → Modify → Payment → Confirmation

the customer could simply tell the agent what they want.

The agent handles the underlying workflow.

This is a fundamental change in software interaction.

What Happens to Traditional Websites?

If Zhang's thesis is correct, customer-facing software could evolve significantly.

Today's model:

Customer → Website/App → Forms → Menus → Support

Future model:

Customer → AI Concierge → Business Systems

The AI becomes the interface.

The underlying website and applications still exist, but the customer may interact with them less directly.

This is one of the most ambitious implications of Decagon's strategy.

Decagon's Enterprise Architecture

The challenge is that AI agents need access to real business systems.

An agent might need to interact with:

CRM
Order management
Billing
Payments
Inventory
Customer databases
Knowledge bases
Identity systems

Decagon therefore isn't merely an AI interface.

It is effectively an orchestration layer connecting AI reasoning with enterprise infrastructure.

Reliability Is the Real Challenge

Customer-service AI cannot simply be impressive.

It has to be reliable.

If an AI gives a slightly wrong answer in a casual conversation, the consequences may be small.

If it incorrectly:

Refunds a transaction
Cancels an account
Changes a reservation
Replaces a credit card

the consequences can be significant.

Decagon therefore emphasizes:

Security
Agent governance
Testing
Observability
Guardrails
Tool permissions
Human escalation

These capabilities are critical to turning AI demos into enterprise software.

The "Agent Engineer"

Zhang has also promoted the concept of the Agent Engineer.

In his March 2026 memo, he describes a future in which building AI agents becomes a distinct engineering discipline.

The role combines:

Software engineering
AI
Product development
Customer understanding
Workflow design

The Agent Engineer isn't simply writing code.

They are designing systems capable of interacting with customers and executing real-world business processes.

Building a Different Type of Engineering Organization

This philosophy influences Decagon's culture.

Instead of separating:

Engineering → Product → Customer Support

the company encourages teams to work directly with customer problems.

That creates a tight feedback loop:

Customer problem → Agent design → Deployment → Real-world data → Improvement

This can accelerate product development dramatically.

Speed as a Competitive Advantage

Zhang has repeatedly emphasized speed.

Decagon's customers have reported being able to build and deploy complex AI-agent workflows much faster than with competing systems.

In June 2026, Decagon said a Fortune 50 customer built and iterated five complex customer-facing workflows in three weeks, after spending nine months attempting to achieve similar results with another AI CX platform.

The company also cited a global airline launching in under three weeks and a major music-streaming platform moving from proof of concept to production in six business days.

These examples illustrate what Decagon sees as one of its key advantages:

time-to-production.

The Importance of Customer Feedback

Zhang's founder philosophy is strongly product-oriented.

In interviews, he has emphasized solving real problems rather than simply building technology because it is technically interesting.

This is a continuation of the philosophy he developed at Lowkey.

The technology can change.

The fundamental question remains:

Does the product solve a problem people care enough about to use?

A Second-Time Founder Advantage

One of Zhang's most important advantages is that Decagon isn't his first company.

Lowkey gave him experience with:

Fundraising
Hiring
Product development
Growth
Company culture
Acquisition
Working with major technology companies

That experience likely helped him avoid some of the common mistakes made by first-time founders.

The difference is visible in Decagon's speed.

The company moved from founding to enterprise customers, major venture rounds, and multibillion-dollar valuation remarkably quickly.

From Consumer to Enterprise

The transition from Lowkey to Decagon is particularly interesting.

Lowkey

Consumer

Gaming

Social

Community

Decagon

Enterprise

Customer experience

Automation

The markets are completely different.

But the underlying founder skill is the same:

Understand user behavior and build a product that changes it.

Angel Investor

Zhang has also become an active angel investor.

His portfolio has included companies such as:

Lovable
Cursor
Listen Labs
Moment
Succinct
Cognition
Pika
Brain Co
Etched
Reducto

His LinkedIn profile describes him as an angel investor across more than 20 startups.

This gives Zhang exposure to a broad cross-section of the AI startup ecosystem.

His AI Investment Thesis

Zhang's investment activity provides another window into his thinking.

Many of the startups associated with his angel portfolio are working on:

AI coding
AI agents
Developer tools
AI infrastructure
Generative media
Enterprise AI

This suggests that he views AI not as a single software category but as a fundamental technological shift affecting many layers of computing.

Zhang's View of the AI Market

In 2026, Zhang has increasingly written about the economics of AI.

His public writing has covered topics including:

AI inference costs
Open-source AI
Enterprise AI adoption
AI agent deployment
The economics of AI software

In July 2026, for example, he argued that falling token prices do not necessarily mean that AI companies' total AI bills will fall, because increased usage can offset lower unit costs.

This reflects his engineering-and-business perspective.

The AI Cost Curve

One of the important economic questions for AI agents is:

Who pays for the inference?

Traditional SaaS has extremely low marginal costs.

AI agents can generate significant inference costs because every customer interaction may require:

Model calls
Reasoning
Tool calls
Retrieval
Verification
Multiple model steps

This makes AI-agent economics fundamentally different from conventional SaaS.

Zhang is therefore building a company where the relationship between:

Revenue → Usage → Inference Cost → Gross Margin

matters enormously.

The Agentic Future

Zhang's long-term vision is not limited to customer service.

He believes AI agents can become a new layer of software that interacts directly with people and businesses.

The progression could look like:

Software applications

↓

AI assistants

↓

AI agents

↓

Autonomous business workflows

↓

AI-native companies

This is why Decagon's potential market extends far beyond customer-support tickets.

Customer Service as the Entry Point

Customer service is the wedge.

But the underlying technology could eventually support other functions:

Sales
Account management
Operations
Commerce
Customer success
Retention
Onboarding
Financial services

Once an AI agent can reliably interact with customers and enterprise systems, the same infrastructure can potentially be applied to many other workflows.

The Competitive Landscape

Decagon competes in one of the hottest categories in enterprise AI.

Its most visible competitor is Sierra, founded by Bret Taylor and Clay Bavor.

The strategic similarities are striking.

Both companies are attempting to:

Build enterprise AI agents
Automate customer service
Integrate with business systems
Move beyond chatbots
Charge for outcomes
Become a new interface between customers and companies

But the competition may ultimately help establish AI agents as a major software category.

Zhang vs. Traditional SaaS CEOs

Traditional SaaS CEOs typically optimize:

Seats × Price × Retention

AI-agent CEOs increasingly need to optimize:

Tasks × Success Rate × Outcome Value × Inference Economics

That requires a different approach to software economics.

Zhang is operating directly inside this transition.

Leadership Style

Zhang's leadership philosophy appears to emphasize:

Speed

Move quickly from idea to production.

Technical Excellence

Treat AI engineering as a core competitive advantage.

Customer Obsession

Work directly with customers and understand their workflows.

Small, High-Performance Teams

Give highly capable employees substantial ownership.

Product Thinking

Build around customer outcomes rather than AI capabilities alone.

Long-Term Ambition

Use customer service as the entry point to a much larger AI-agent platform.

Jesse Zhang – Key Facts

Full Name: Jesse Zhang
Current Position: Co-Founder & CEO, Decagon
Company: Decagon
Founded: 2023
Co-Founder: Ashwin Sreenivas
Headquarters: San Francisco
Previous Startup: Lowkey
Lowkey: Acquired by Niantic in 2021
Education: Harvard University
Degree: Computer Science
Graduation: Three years
Previous Experience: Google, Citadel, Hudson River Trading and others
Decagon Series C: $131M
2025 Valuation: $1.5B
Series D: $250M
2026 Valuation: $4.5B
Estimated 2026 Annualized Revenue: ~$100M
Core Product: Enterprise AI agents
Primary Market: Customer experience / customer service
Channels: Chat, email, voice
Key Product Concept: Agent Operating Procedures (AOPs)
Known For: AI agents, enterprise automation, customer experience, product-led AI
Angel Investments: 20+ startups reported on his LinkedIn profile.

Jesse Zhang's Entrepreneurial Legacy

Jesse Zhang's career illustrates how quickly the path from young founder to major technology CEO can now happen in the AI era.

He began with Lowkey, a consumer gaming and social company.

The startup was acquired by Niantic in 2021.

Instead of remaining inside a large technology company, Zhang returned to entrepreneurship and identified a much larger technological shift:

AI agents.

With Ashwin Sreenivas, he founded Decagon in 2023 around a simple but powerful idea:

Customer service shouldn't require a human to manually execute every step of a workflow.

The first generation of AI customer service software focused primarily on conversation.

Zhang is betting on something much bigger:

Conversation → Reasoning → Action → Resolution

That distinction has helped Decagon grow at extraordinary speed.

The company reached a $1.5 billion valuation in 2025, raised another $250 million in early 2026, and reached a $4.5 billion valuation in less than three years from founding.

By 2026, Decagon's vision had expanded beyond reactive customer support.

Its agents can work across voice, email and chat; proactive agents can initiate interactions; and the company's AOP framework gives businesses a way to define how their agents behave.

Zhang's broader thesis is even more ambitious.

He believes that every major brand could eventually have an AI agent serving as its primary interface with customers.

If that happens, the traditional relationship between customers and businesses could change fundamentally.

Today:

Customer → Website → App → Forms → Support Agent

Tomorrow:

Customer → AI Concierge → Business Systems → Outcome

For CEO-Watch, Jesse Zhang is particularly interesting because he represents the second-generation AI founder.

He isn't simply building an AI feature.

He is building an organization around the assumption that AI agents will become a new software layer.

His previous experience as a young consumer founder, his technical background, his acquisition by Niantic, and his rapid transition into enterprise AI give him an unusually broad perspective.

The biggest question surrounding Zhang now is no longer whether Decagon can build a successful AI customer-service company.

The company has already demonstrated substantial commercial traction.

The much bigger question is whether Decagon can become the agentic infrastructure layer through which businesses interact with their customers.

If Zhang succeeds, Decagon could help define a fundamental shift in enterprise software:

from software that humans operate to software that acts on behalf of humans.

That is the bet making Jesse Zhang one of the most important young CEOs to watch in AI and enterprise SaaS.