20VC: Why 90% of Founders Build Startups Wrong | Why AI Growth Rates are Sustainable & Remote Work is BS and the AI Talent War | Competing with Brett Taylor and Sierra: Who Wins the Customer Service War with Jesse Zhang, Decagon - The Twenty Minute VC (20VC) | Ryan Randels | Mobile & SaaS Strategy | Agentic AI

Full Title

20VC: Why 90% of Founders Build Startups Wrong | Why AI Growth Rates are Sustainable & Remote Work is BS and the AI Talent War | Competing with Brett Taylor and Sierra: Who Wins the Customer Service War with Jesse Zhang, Decagon

Summary

This episode features Jesse Zhang, co-founder and CEO of Decagon, a conversational AI platform for customer experience. The discussion covers effective startup building strategies, the sustainable growth of AI, the value of in-person work, and the competitive landscape in AI-driven customer service. Zhang shares insights from his previous entrepreneurial experience and his journey with Decagon, emphasizing customer discovery and the transition from traditional software to AI solutions.

Key Points

Math Olympiad background can foster strong reasoning and problem-solving skills beneficial for founders, though often historically channeled into research or trading rather than entrepreneurship.
Early startup success, like the acquisition of Loki, can ground founders for future endeavors, but over-intellectualizing and building for perceived trends rather than customer needs can lead to demoralizing failures.
The "right way" to build companies in 2025 involves deep customer discovery and iterating through practice ("reps") rather than solely relying on hype or theoretical knowledge, as true market fit is found through execution.
Execution can trump market selection in early stages, as effective discovery helps identify viable markets, as demonstrated by Decagon's shift to focusing on conversational AI for customer service after initial explorations into broader AI applications.
Over-promising and selling ahead of time, as seen with Harvey, can be a viable strategy in some markets where momentum and early adoption by key players drive success, but in broader markets, demonstrating tangible ROI quickly is crucial.
Decagon competes with larger players like Salesforce in the customer service AI space, balancing their agility in product development against Salesforce's extensive distribution channels, while acknowledging a "land grab" due to current market excitement.
The zero-to-one phase for Decagon, executed by two co-founders, was accelerated by deep customer discovery, leading to a rapid ramp to $1M ARR, emphasizing the importance of customer validation over mere excitement for a product.
Brand names of venture investors can be important for talent attraction and customer validation, though some argue that a hire's focus solely on investor brand might indicate they are not the right fit.
While brand VCs offer validation, their ability to directly accelerate a startup's path to product-market fit (PMF) is debatable, with platform teams being more valuable after initial traction.
The AI market is largely starting fresh, benefiting startups with AI-native approaches over established players who may be hampered by legacy systems and customer dependencies, creating an advantage for agile newcomers.
AI's true unlock is democratizing complex tasks for non-technical users through natural language interfaces, enabling rapid iteration and customizability, as seen with Decagon's "agent operating procedures" (AOPs).
The transition from software spend to human labor budgets is occurring in AI, with application-layer solutions justifying higher costs by offering significant ROI and solving problems more completely than traditional software tools.
Customer service AI is a market with true PMF because it effectively shifts spend from human labor to software, offering a clear ROI, unlike some other AI tools where commoditization and low switching costs limit pricing power.
Enterprise sales, especially for AI solutions, involve more than just product delivery, encompassing implementation, customization, and strategic alignment, which creates higher switching costs and value capture compared to product-led growth (PLG) models.
The AI industry's future likely involves multiple winners rather than a single dominant player due to the lack of strong network effects between customers and varying company needs, though switching costs will remain significant.
Contextual memory and the "system of intelligence" are key to customer lock-in in AI, with proprietary business logic and configurations being harder to port than simple data or conversation logs.
Future AI interaction may move beyond current prompting methods towards more intuitive, example-based learning and agent shadowing, reflecting how humans learn.
"Clock speed" – the speed of thought and learning – is highly valued by Decagon for all roles, including sales and marketing, over specific industry experience.
Over-indexing on forward-deployed engineering (FDE) might be a mistake in use cases where a product-first approach is more scalable, as customer needs can be consistent across the board.
Embracing stress as an advantage rather than solely mitigating it can lead to a more exciting and productive life, reflecting a perspective that current perceived hardships are minor compared to historical struggles.
Decagon focuses on building a winning culture by ensuring employees are learning, earning, and part of a successful, fast-moving team.
Decagon's biggest internal debate is the balance between servicing immediate customer needs and investing in long-term product development, a common challenge for rapidly growing companies.
Hiring engineers is crucial for AI companies, and Decagon prioritizes offering a unique culture and opportunity for career advancement over solely competing with the cash of larger tech giants.
Decagon's leadership style is intense and geared towards winning, with a future goal of incorporating more nuanced "softness" alongside this intensity.
While AI can be transformative, many enterprise use cases are not yet fully realized due to the technology not being "good enough" or the problem's nature not being amenable to current AI capabilities.
Salesforce's distribution is the most desirable poached asset from a competitor due to its vast reach and established customer relationships.
The success of Salesforce's Agent Force is questioned due to potential fragmentation and difficulty moving fast with a large customer base, contrasting with Decagon's product-forward approach.
The transition from traditional enterprise software sales, characterized by heavy configuration and longer implementation, differs from Decagon's more product-led, agile approach, appealing to different customer segments.
Decagon prioritizes hiring engineers over exclusive access to new AI models, believing that strong talent and platform development around existing models are more crucial for competitive advantage.
In-person work is preferred for its productivity and the faster flow of ideas, contrasting with remote work's limitations during COVID.
Having a team predominantly in their 30s, with leadership married young, contributes to a culture of hard work and shared ambition focused on winning.
The biggest internal debate at Decagon is the allocation of resources between immediate customer success and long-term product innovation.
The AI market is unlikely to be a winner-take-all scenario, with multiple strong players expected due to the nature of enterprise software.
The primary challenge for AI B2B companies in the next 12-18 months will be hiring talent, not necessarily competing directly with the cash reserves of giants like OpenAI or Anthropic.
High valuations, while achievable, can create demotivating goalposts for teams, complicate hiring with unrealistic equity expectations, and limit future optionality, potentially leading to "zombie company" status.
The "job displacement" question is a frequently asked and somewhat repetitive topic in AI discussions.
The most significant change in AI's operational impact will likely be in sales, with current AI applications for sales productivity being basic and having potential for more creative and impactful uses.
Anthropic is favored over OpenAI in a hypothetical investment scenario due to the belief that the AI market won't be a winner-take-all, Anthropic's strong work in coding agents, and its promising team, despite OpenAI's consumer dominance.
A key reflection for improvement is the tendency for leaders to get caught in the weeds of specific deals rather than focusing on high-level strategy.
The core advantage of AI solutions comes from building effective platforms and orchestration around models, rather than solely relying on the marginal improvements of newer models.
The future of AI interaction may involve more learning from examples and less explicit prompting, mimicking human learning processes.

Conclusion

Effective startup building hinges on deep customer discovery and continuous execution, with a focus on solving real problems rather than chasing market hype.

The AI revolution offers a unique opportunity for agile startups to innovate by building from scratch, unburdened by legacy systems, and democratizing complex tasks through intuitive interfaces.

Long-term success in AI will likely come from building robust platforms and orchestrating capabilities around underlying models, coupled with a strong culture of winning and continuous learning.