Invest Like the Best – EP.443 – Jesse Zhang: Building Decagon

Podcast Host: Patrick O'Shaughnessy
Guest: Jesse Zhang, Co-Founder & CEO, Decagon
Date: October 6, 2025

Episode Overview

This episode dives into the world of AI-powered customer service with Jesse Zhang, CEO and co-founder of Decagon, one of the fastest-growing companies in the conversational AI space. Patrick and Jesse dissect Decagon's journey in finding product-market fit, the evolving landscape of AI startups, strategic choices around proprietary models and agent deployment, and the intense culture necessary to compete in today’s AI battlegrounds. Jesse shares his playbook for uncovering enterprise demand, lessons from his own founder journey, and sharp takes on industry-wide topics from talent wars to future business models.

Key Discussion Points & Insights

1. The Culture of Winning, Competition, and Intensity

• Office Motto & Culture
  Jesse describes Decagon's office motto: “There’s no challenge that can’t be overcome and there’s no enemy that can’t be defeated.”
  “We have a very competitive team. Everyone wants to win... It really fits our culture.” [05:13]
  Inspiration partly comes from Huawei’s legendary culture.

• Aggression & Competition in Modern AI Startups
  Jesse examines how intense, almost combative cultures have become normalized—and sought after—in AI and tech startups:

  “In this generation, it almost attracts a specific demographic... people that grew up in competitive environments, academics or whatever. A lot of the people around my age are doing startups now, and people are doing quite well.” [06:53]

• Math Contests as a Talent Funnel
  Both Jesse and Scott Wu (Cognition) come out of elite math competition circles, shaping their objective, competitive, and community-driven worldviews.

  “How well you’re doing is fairly objective... there’s constant motivation to improve. That’s quite nice.” [08:44]

2. Decagon’s Approach to Product-Market Fit and Customer Validation

• Systematic Solution Testing
  Jesse details a rigorous, almost sales-like process asking prospects not just for their problems, but for how much they’d pay to solve them:

  “When you talk to a potential customer... they actually don’t mind answering questions such as, okay, if we built this for you, how much would you pay for it? Would your boss need to approve it?” [13:38]

• Deep Customer Probing Yields Real Signal

  “Most of the time at the end of this exercise you’re like, okay, glad I didn’t pursue this further because that would have been a waste of time. ...But it puts the customer in the same frame of mind as you.” [16:32]

• Finding the Breakout Use Case: Customer Service
  Decagon’s conviction materialized because the revenue willingness for customer service AI was 10x any other idea.

  “It was clearly the thing because if you tallied up the amounts that people said, this was probably an order of magnitude more than anything else.” [20:25]

3. Why Customer Support & Code Generation Are the AI Sweet Spots

• Enterprise Desires Clear ROI and Containment

  “The ROI is really easy to justify... [and] it’s very easy to go live... you already have your call center, your telephony stack, so you just connect to it.” [21:39]

• Replacement vs. Augmentation Spectrum
  Jesse frames AI agents as attacking both ends of the labor spectrum: replacing the most routine/low-cost customer service tasks, and augmenting the most expensive (developers).

  “AI use cases will start eating the spectrum from both ends. For customer service, it’s more of the replacement sense; for coding, it’s augmentation.” [23:56]

• Minimizing Risks When Deploying AI Agents
  Rapid rollout is possible due to robust escalation paths and metrics-driven monitoring:

  “Even within a week... What is the resolution rate? Is that what we expect? What is the customer satisfaction?” [28:56]

• Failure Modes and Surprise Wins
  Early funny/awkward failures included misinterpreting customer intentions, e.g., responding cheerfully to unexpected behavior:

  “The agent was like, oh my God, that’s so awesome that you’re thinking about doing something nice for the community.” [29:57] But major wins came from reducing requests to “speak to a human” and building genuine customer trust. [32:19]

4. AI Agents: Execution, Voice, and the Future Customer Experience

• Emerging Importance of Voice-to-Voice AI

  “The biggest frontier right now is voice models... The bar is very high. Uncanny Valley is quite large.” [33:30] Voice-to-voice models enable nuance, emotion, and low-latency flows, but bring technical hurdles like higher hallucination rates: “Probably like 8x higher or something like that.” [36:25]

• Agent Capabilities Today and Tomorrow
  Customer interactions handled by Decagon agents range from simple FAQ to personalized problem-solving and intricate account actions. [37:22] The eventual vision? Agents as full digital concierges, the unified “front door” to a business spanning all customer touchpoints. [44:37]

5. Data Moats, Proprietary Models, and Business Model Evolution

• Data as a Learning and Growth Engine Decagon aims to “read every conversation and extract whatever info you want from it,” allowing “the agent to improve automatically over time” and creating strong data-driven moat effects. [39:35]

• Fine-Tuning, Model Choices, and Enterprise Application Moats
  Jesse sees a trend towards using a blend of large and fine-tuned smaller models for efficiency and specialization:

  “Nowadays we’re seeing much more of that because applications have gotten more mature.” [59:11]

• ChatGPT 'Wrappers'—Myth and Reality Jesse distinguishes thin wrappers from full enterprise platforms:

  “Most of the time that’s not the case. Especially when you get into agents, an agent is not just a model—you have to design it, put in guardrails, teach it. And that’s where the software layer comes in.” [69:54]

6. Investing, Fundraising, and Talent Wars

• AI Startup Fundraising Mania

  “It definitely feels like there’s maybe a little bit too much excitement right now on the AI side. It just seems way too easy to raise money.” [49:39]

  The best investors demonstrate helpfulness pre-investment, with a focus on customer insight, strategic thinking, and culture fit. [52:12–54:03]

• Team-Building and Recruiting Decagon’s culture is marked by intensity, hard work, and strong in-person collaboration:

  “Everyone that has joined Decagon... they want to work hard, be around other smart, motivated people.” [55:14] Hiring involves “swarming” top candidates, engaging not just the person but their families and motivations. [57:10]

• Talent Wars in Depth While the hottest research wars happen at the model layer, Decagon still contends for elite applied research and engineering talent, and has opened in-person offices in multiple cities to access more pools. [57:10–58:25]

7. Commercial Focus and Founder Perspective

• Jesse’s Commercial-First Mindset Compared to peers who may chase technology for its own sake, Jesse is laser-focused on commercial value and practical wins:

  “I generally lean a lot more towards commercial elements of every idea... I think you should really index on the commercial side.” [73:43]

• Milestone Setting and Team Motivation The team rallies around clear, tangible short-term goals—sometimes as simple as custom jackets for revenue wins.

  “Last year for our revenue milestone, we told everyone we’d get them super nice jackets... it just creates this, hey, we’re working towards these jackets.” [74:56]

Notable Quotes & Memorable Moments

• On Competitive Culture:
  “Words like ‘defeated’... ‘violence’... ‘aggression’—these are not words that were being used three years ago or four years ago. In fact, if you used them, it was a big problem... That has completely shifted.”
  —Patrick O’Shaughnessy [06:05]

• On Product-Market Fit Discovery:
  “If we built this for you, exactly how much would you pay for it? Would your boss need to approve it or your boss's boss? How would the entire organization think about ROI?”
  —Jesse Zhang [13:38]

• Philosophy on Running at Losses:
  “What you’re really optimizing for right now is just quality and growth. The optimization will always come later on.”
  —Jesse Zhang [66:32]

• On Voice AI Progress:
  “If you really want to make it indistinguishable from a human, you have to do voice-to-voice, or you have to at least take into account the voice.”
  —Jesse Zhang [35:17]

• On ChatGPT Wrappers:
  “If you have enough software built around the models, then that’s where you can actually almost capture the most value.”
  —Jesse Zhang [68:32]

• On Long-Term Vision:
  “Eventually this becomes the front end for the business... In the limit... the agent could be the only thing most users interact with.”
  —Jesse Zhang [44:37]

Timestamps for Key Segments

| Timestamp | Topic | |-------------|-------------------------------------------------------------------------------| | 05:13 | Decagon’s culture and defining office motto |
| 06:53 | Competitive dynamics in today’s AI startup scene |
| 08:44 | Math contests as a foundation for startup success |
| 13:38 | Decagon’s approach to customer discovery and validation |
| 16:32 | Structuring deep, high-signal client conversations |
| 20:25 | Deciding on AI customer service as the focus use case |
| 21:39 | Why customer support is so ripe for AI disruption |
| 23:56 | Comparing customer support vs. code-as-agent use cases |
| 29:57 | Early product failure story: the “homeless people” customer interaction |
| 32:19 | Surprising wins in customer trust and agent adoption |
| 33:30 | State and frontier of voice AI |
| 39:35 | Data flywheel and automatic product improvement |
| 41:43 | Frameworks for enterprise AI opportunity assessment |
| 49:39 | Fundraising climate and investor behavior in AI boom |
| 55:14 | Decagon’s team culture and recruiting philosophy |
| 57:10 | Talent wars in AI—inside perspective |
| 59:11 | Strategic model decisions: open source, fine-tuning, and long-term power |
| 69:54 | On “wrappers”—thin apps vs. true enterprise tools |
| 74:56 | Setting team milestones and motivation strategies |
| 78:28 | Economic scalability and forward deployed engineering models |
| 80:00~ | The kindest thing Jesse’s parents did for him (personal closing reflection) |

Final Reflections

This conversation offers an inside look at the DNA of a generational AI company: a cocktail of data-obsessed rigor, intense competition, methodical validation, and relentless commercial focus. Decagon’s story is both a playbook for enterprise AI success and a reality check on what it takes—culturally, technically, and commercially—to win in the hottest market of the decade.

Listen if:
You’re building, investing, or leading in the AI or SaaS world, and want to learn how the very best think about customer value, company culture, enterprise adoption, and sustainable advantage.